> Bob needs a new computer for his job.... In order to obtain a new work computer he has to create a 4 paragraph business case explaining why the new computer will improve his productivity.
> Bob’s manager receives 4 paragraphs of dense prose and realises from the first line that he’s going to have to read the whole thing carefully to work out what he’s being asked for and why. Instead, he copies the email into the LLM.... The 4 paragraphs are summarised as “The sender needs a new computer as his current one is old and slow and makes him unproductive.” The manager approves the request.
"LLM inflation" as a "bad" thing often reflects a "bad" system.
In the case described, the bad system is the expectation that one has to write, or is more likely to obtain a favorable result from writing, a 4 paragraph business case. Since Bob inflates his words to fill 4 paragraphs and the manager deflates them to summarise, it's clear that the 4 paragraph expectation/incentive is the "bad" thing here.
This phenomenon of assigning the cause of "bad" things to LLMs is pretty rife.
In fact, one could say that the LLM is optimizing given the system requirement: it's a lot easier to get around this bad framework.
The 4-paragraph business case was useful for creating friction, which meant that if you couldn't be bothered to write 4 paragraphs you very likely didn't need the computer upgrade in the first place.
This might have been a genuinely useful system, something which broke down with the existence of LLMs.
The only definitively non-renewable resource is time. Time is often spent like a currency, whose monetary instrument is some tangible proxy of how much time elapsed. Verbosity was an excellent proxy, at least prior to the advent of generative AI. As you said, the reason Bob needs to write 4 paragraphs to get a new PC is to prove that he spent the requisite time for that computer, and is thus serious about the request. It’s the same reason management consultants and investment bankers spend 80+ hours a week working on enormous slide decks that only ever get skimmed by their clients: it proves to the clients that the firm spent time on them, and is thus serious about the case/deal. It’s also the same reason a concise thank-you note “thanks for the invite! we had a blast!” or a concise condolence note “very sorry for your loss” get a lot less well-received than a couple verbose paragraphs on how great the event was or how much the deceased will be missed, even if all that extra verbiage confers absolutely nothing beyond the core sentiment. (The very best notes, of course, use their extra words to convey something personally meaningful beyond “thanks” or “sorry.”)
Gen-AI completely negates meaningless verbosity as a proxy of time spent. It will be interesting to see what emerges as a new proxy, since time-as-currency is extremely engrained into the fabric of human social interactions.
There's an important asymmetry here: it takes a lot of to weave an intricate pattern, but much less time to assess and even appreciate it. The sender / suitor pays significantly more than the receiver / decider.
There are some parallels to that in compression and cryptography, but they are rather far-fetched.
This is the sort of workplace philosophising that I hate the most. Employees aren't children. They don't need to have artificial bullshit put up in between them and what they need, the person approving just needs to actually pay attention.
If someone wants a new computer they should just have to say why. And if it's a good reason, give it to them. If not, don't. Managers have to manage. They have to do their jobs. I'm a manager and I do my job by listening to the people I manage. I don't put them through humiliation rituals to get new equipment.
People want new computer because new hire Peter got a new one.
People want new computer because they just had lunch with a friend that works in a different company and got a new computer and they just need to one up him next time they go to lunch.
That is why I am not going to just give people computers because they ask. Worst crybabies come back because they „spilled coffee” on perfectly fine 2 years old laptop.
Wooo I used to think this was how managers work and just ... Was inevitable. I'm so glad to actually be a manager now, because no, it's not. You don't call people who spill coffee, (have you never spilled anything?) crybabies. This is a bad manager.
I can say when I wanted to move my desk from one place to another I had to write up the "business justification" (I was already working from both offices on different days and still am, it was a change on paper)
so I'm sure there's large corps that do this for everything. probably ones where you're not asking your manager, but asking finance or IT for everything
The problem is, I'm a verbose writer and can trivially churn out 4 paragraphs - another person is going to struggle. The friction is targeting the wrong area: this is a 15 minute break for me, and an hour long nightmare for my dyslexic co-worker.
Social media will give you a good idea what sort of person enjoys writing 4 paragraphs when something goes wrong; do you really want to incentivize that?
I love this article for how it gets the thinking, and I love your response.
I've been aware of similar dynamic in politics, where the collective action/intelligence of the internet destroyed all the old signals politicians used to rely on. Emails don't mean anything like letters used to mean. Even phone calls are automated now. Your words and experience matter more in a statistical big data sense, rather than individually.
---
This puts me in sci-fi world-building mode, wondering what the absurd extension is... maybe it's just proving burned time investment. So maybe in an imagined world where LLMs are available to all as extensions of thought via neural implant, you can't be taken seriously for even the simplest direct statements unless you prove your mind sat and did nothing (aka wasted it's time) for some arbitrary period of time. So if you sat in the corner and registered inactive boredom for 2h, and attached a non-renewable proof of that to a written word, then people would take your perspective seriously, because you expended (though not "gave") your limited attention/time to the request for some significant amount of time
I'm a consultant, and have been for 20 years now (except for a 2 year stint as an employee for the benefits and relocation assistance to move my family from Texas to NYC).
At the NYC employment job, I was on a 2 year upgrade cycle being a software developer. Just whatever the current year Dell corpo laptop was. No procurement procedures, just IT got 2 year replacements, our laptops went to the non IT workforce after reimaging.
As a consultant I usually bring my own device, and my laptops are usually WAY more capable since I will run each client on their own VM - makes it easy to delete the client when the contract is up. But I've had one client who did not allow BYOD, and any billed work had to be done on their hardware. That was fine, except that the desktop I was given was already a 12 year old dual core non-hyperthreading CPU that wasn't meant for developers even when it was built. I begged and pleaded for 6+ months for me to either bring in my own hardware they could image now and wipe at the end of the contract, or to please buy me a PC from this decade.
It took 3 years to get the budget approval for a $2000 tower, roughly the equivalent of 15 hours of pay. The thing that finally pushed it over the edge, was that my PC could not handle Teams + Visual Studio at the same time, and manager couldn't handle that he couldn't watch me program.
All of that to say I doubt these non-data-driven organizations are basing these decisions on anything other than micromanagement. Nothing to do with measured or assumed productivity, nothing to do with costs, so all I can think is they have to be a "decision maker" on all aspects.
I never understood why no company I worked at could hire someone just to manage the kitchen instead of paying engineers massive salaries to argue about unloading the dishwasher.
Sounds like your company needs to spend some time working on their onboarding process. Probably more effective than just hoping no one ever needs a new computer.
Whether it's an additional $100/week for base salary or $100/week for overhead Bob is always going to be after another $100/week regardless what his current cost is. If Bob wants to manage the overheads directly (most don't) like salary costs then he probably wants more of a contractor type position.
As to the content of the letter, the 4 paragraphs are supposed to be "these are reasons I think were missed and why it'll cost more to not correct it" not just "I put effort to write 4 paragraphs of stuff" friction alone.
Having run a short stint as an internal IT manager at an IT focused company... it's astounding how many non-standard/out-of-cycle laptop request are actually either basic user error (even for the most brilliant technical employees) or basic IT systems problems (e.g. poorly tested management/security tool changes eating up performance/battery in certain configurations) that a new laptop won't actually solve. E.g. reports of "my battery runs out in 2 hours and my IM is dog slow" but they are on an M1/M2 MacBook Pro and probably wouldn't notice if they got an M1 or M4 MacBook back as their issue isn't actually the hardware. When someone writes an email or ticket explaining why their use case just wasn't accounted for it's generally pretty obvious they really do need something different.
> Bob’s manager receives 4 paragraphs of dense prose and realises from the first line that he’s going to have to read the whole thing carefully to work out what he’s being asked for and why. Instead, he copies the email into the LLM.... The 4 paragraphs are summarised as “The sender needs a new computer as his current one is old and slow and makes him unproductive.” The manager approves the request.
Engaging with why we might actually want inflation of text:
1) For pedagogical or explanatory purposes. For example, if I were to write:
> ∀x∈R,x^2≥0
I've used 10 characters to say
> For every real number x, it's square is greater than or equal to zero
For a mathematician, the first is sufficient. For someone learning, the second might be better (and perhaps as expansion of 'real number' or that 'square' is 'multiplying it by itself').
2) To make sure everything is stated and explicit. "He finally did x" implies that something has been anticipated/worked on for awhile, but "after a period of anticipation he did x" makes it more clear. This also raises the question of who was anticipating, which could be made explicit too.
As someone who spends a lot of time converting specifications to code (and explaining technical problems to non-technical people), unstated assumptions are very prevalent. And then sometimes people have different conceptions of the unstated assumption (i.e. some people might think that nobody was anticipating, it just took longer than you'd expect otherwise).
So longer text might seem like a simple expansion, but then it ends up adding detail.
I definitely agree with the authors point, I just want to argue that having a text-expander tool isn't quite as useless as 'generate garbage for me'.
Yes, because generators generate at the token level, which is technically smaller than an individual word. They can easily generate unique sentences, and for example transfer learning allows them to apply knowledge obtained from some other training data to new domains.
The idea that generators are some sort of parrot is very outdated. The 2021 paper that coined the term "stochastic parrot" was already wrong when it was published.
> Yes, because generators generate at the token level, which is technically smaller than an individual word. They can easily generate unique sentences, and for example transfer learning allows them to apply knowledge obtained from some other training data to new domains.
Sure. But can they read the original author's mind, and therefore generate the right unique sentence that expresses the actual intent?
Sure, but in the case of "he finally did X" without context passed in, how does the llm determine if it would be expanded as "this was a very anticipated change" or if the author is frustrated at how long it took? If the nuanced meaning isn't there in the input context.
obviously it can generate the longer message, but is it going to go look up what the sentence refers to and infer extra meaning automatically...?
The important part is the GDP is now increased because of the cost of energy and additional hardware needed expand and then compress the original data. Think of the economic growth all these new hassles provide!
The older I get the more concise I find myself (which is not to say I'm actually concise, as my comment history will demonstrate), but LLM's have really driven home just how much noise day to day communication involves. So much filler text.
It still surprises me when I see non-technical enthusiasts get excited about LLMs drafting almost useless copy or email or whatever. So much garbage text no one reads but has to be written for some reason. Its weird.
"I wrote this mail slightly longer because I didn't have time to make it short" - someone famous
When writing something I want people to read, I always take time at the end to make it shorter - remove distracting sentences, unnecessary adjectives and other noise. Really works wonders for team communication.
>When writing something I want people to read, I always take time at the end to make it shorter - remove distracting sentences, unnecessary adjectives and other noise.
This is a good advice. How can I do it when talking? I often talk too much saying little, often loosing Listener's attention in the process.
On one side you have people using LLMs to fluff a sentence in to an essay. And on the receiver side they are hitting a button to AI summarise it back to a sentence.
An LLM is effectively a compressed model of its input data.
Inference is then the decompression stage where it generates text from the input prompt and the compressed model.
Now that compressing and decompressing texts is trivial with LLMs, we humans should focus - in business at least - on communicating only the core of what we want to say.
If the argument to get a new keyboard is: "i like it", then this should suffice, for inflated versions of this argument can be trivially generated.
What I hate about this is that often a novel and interesting idea truly needs extra space to define and illustrate itself, and by virtue of its novelty LLMs will have substantially more difficulty summarizing it correctly. But it sounds like we are heading to a medium-term where people cynically assume any long email must be LLM-generated fluff, and hence nothing is lost by asking for an LLM summary.
Not to be overly snide, but I can imagine that almost every person who writes long, tedious emails that wax on and on thinks they have something novel and interesting that truly needs extra space. Also, most novel things are composed of pedestrian things, which LLMs have no issue summarizing sufficiently.
Maybe you can provide an example where this case would occur, and maybe some indication how often you think this would occur.
It seems to be a big problem for AI-generated summaries of scientific research, where the top-performing AIs ignores key details about 25% of the time, and often much worse. The problem is exactly what I said: the research has a new idea which is (by necessity) not going to be in the pretraining data, so the AI ignores or misstates it. I don't expect this problem to be solved anytime soon, it is inherent to how 2020s artificial neural networks work.
> If the argument to get a new keyboard is: "i like it", then this should suffice
This seems like exactly what LLMs are supposed to be good at, according to you, so why don't they just near-losslessly compress the data first, and then train on that?
Also, if they're so good at this, then why are their answers often long-winded and require so much skimming to get what I want?
I'm skeptical LLMs are accurately described as "near lossless de/compression engines".
If you change the temperature settings, they can get quite creative.
They are their algorithm, run on their inputs, which can be roughly described as a form of compression, but it's unlike the main forms of compression we think of - and it at least appears to have emergent decompression properties we aren't used to.
If you up the lossy-ness on a JPEG, you don't really end up with creative outputs. Maybe you do by coincidence, and maybe you only do with LLMs - but at much higher rates.
Whatever is happening does not seem to be what I think people typically associate with simple de/compression.
Theoretically, you can train an LLM on all of Physics, except a few things, and it could discover the missing pieces through reasoning.
Yeah, maybe a JPEG could, too, but the odds of that seem astronomically lower.
If you don't design your compressor to output data that can be compressed further, it's going to trash compressibility.
And if you find a way to compress text that isn't insanely computationally expensive, and still makes the compressed text compressible by LLMs further - i.e. usable in training/inference? You, basically, would have invented a better tokenizer.
A lot of people in the industry are itching for a better tokenizer, so feel free to try.
The inverse of this is "AI Loopidity" where we burn cycles inflating then deflating information (in emails, say, or in AI code that blows up then gets reduced or summarized). This often also leads to weird comms outcomes, like saving a jpg at 85% a dozen times.
I consider inflation a double insult. (https://ruudvanasseldonk.com/2025/llm-interactions) It says "I couldn't be bothered to spend time writing this myself, but I'm expecting you to read all the fluff."
> That we are using LLMs for inflation should not be taken as a criticism of these wonderful tools. It might, however, make us consider why we find ourselves inflating content. At best we’re implicitly rewarding obfuscation and time wasting; at worst we’re allowing a lack of clear thinking to be covered up. I think we’ve all known this to be true, but LLMs allow us to see the full extent of this with our own eyes. Perhaps it will encourage us to change!
Yeah, this is the problem. Wealth distribution stopped working sometime in the late 20th century and we're fighting each others for competitive advantages. That's the core of this phenomenon.
No one needs containers full of baby sized left shoes, but proof of work must be shown. So the leathers must be cut and shoes must be sewn, only to be left in the ever growing pile in the backyard. That's kind of wrong.
Long documents in business contexts that get summarized and go mostly unread are the byproduct of a specific and common level of trust and accountability in those contexts: people don't believe someone has done enough critical thinking or has a strong enough justification for a proposal unless they've put it on the page, but if it is on the page, it's assumed that it does in fact represent critical thinking and legitimate justification.
If trust was higher, shorter documents would be more desirable. If trust was lower, or accountability higher, summarization would be used a lot more carefully.
LLMs haven't changed anything in this regard except that they've made it extremely easy to abuse trust at that specific level. The long-term result will be that trust will fall in the general case, and people will eventually become more careful about using summarization. I don't think it will be long before productized AI used in business contexts will be pretrained/fine-tuned to perform a basic level of AI content detection or include a qualitative measure of information density by default when performing summarization.
Huh, this post is not what I thought it would be! Even after the first two paragraphs!
There's a line of thought which states that intelligence rhymes with compression: Identifying patterns allows better prediction, enables better compression of the data.
However, internally, LLMs typically do the opposite: Tokenization and vectorization multiply the bit rate of the input signal. Chain of thought techniques add a lot of extra text, further increasing the bit rate.
My PM said they’d written a bunch of tickets for a project yesterday morning that we hadn’t fully scoped yet. I was pleasantly surprised because I can’t complain if they are going to get ahead of things and start scaffolding tickets.
Of course when I went to read them they were 100% slop. The funniest requirement were progress bars for actions that don’t have progress. The tickets were, even if you assume the requirements weren’t slop, at least 15 points a piece.
But ok maybe with all of these new tools we can respond by implementing these insane requirements. The real problem is what this article is discussing. Each ticket was also 500-700 words. Requirements that boil down to a single if statement were described in prose. While this is hilarious the problem is it makes them harder to understand.
I tried to explain this and they just said “ok fine rewrite them then”. Which I did in maybe 15min because there wasn’t actually much to write.
At this point I’m at a loss for how to even work with people that are so convinced these things will save time because they look at the volume of the output.
Ask an llm for a project plan and they’ll happily throw dates around for each step, when they can’t possibly know how long it will take.
But project plan dates have always been fiction. Getting there faster is an efficiency win.
That said I’ve found that llms are good as interrogators. If used to guide a conversation, research background information and then be explicitly told to tersely outline the steps in something I’ve had very good results.
The date/week estimations in plans are especially funny when you work with an agent and it spits that out. "Week 1, setting up the structure" - uh, no, we're going to do this in 10 minutes.
> At this point I’m at a loss for how to even work with people that are so convinced these things will save time because they look at the volume of the output.
The same way, presumably, that one used to work with people who would say things like "just look how much code this template system generates for us!" unironically.
The software requirements phase is becoming increasingly critical to the development lifecycle and that trend will continue. I have started writing very short tickets and having claude code inflate them, then I polish those. I often include negative prompts at this point so claude may have included "add a progress bar for xyz" and i simply add "do not" in front of those things that do not make sense. The results have been excellent.
Why? As a frequent consumer of software I love progress bars and I know many others do too. Are you mad other people get paid highly or mad about progress bars? or some combination of both?
If progress bars are actually showing information I want, like how much time is left on the operation, they are great! I suspect most are just garbage though.
The only acceptable response to obvious AI slop - unless it's it's clear it's been heavily reviewed and updated - is to put it back into the AI and ask it for a 1 paragraph summary and work off of that.
As Koffiepoeder suggests, since the vast majority of content on my site is static, I only have to compress a file once when I build the site, no matter how many people later download it. [The small amount of dynamic content on my site isn't compressed, for the reason you suggest.]
Depends on efficiency of compression I guess. If X bytes of data takes N time to transmit, and each slice of N takes Y CPU cycles during transmission, how many Y must your compression algorithm use, and how low must it lower N, in order to be more efficient from a CPU utilisation perspective? assumedly there's an inflection point from CPU use perspective, maybe a point that's impractical to achieve? I'm just vibe-thinking-out-loud.
It depends on where the bottleneck is. If it’s in network packet size, it would help serve more clients. At the expense of more CPU needed to decode/encode the data. If you’re serving large files and you have headroom in hardware it’s totally normal.
LLM Inflation is an interesting choice of terminology. Like with many things in our contemporary society there is a temptation to assign absolute, quantitative value to everyday concepts but realistically we should know this to be a fallacy. Many concepts actually have no "on paper" value but still manifest significant social value. Marketing is the typical example, and yet we don't refer to advertising as inflation (though maybe we should).
This concept probably applies to lots of work in the "AI" space right now. The idea of using huge amounts of compute to generate lifelike voices for LLMs comes to mind as being recently maligned (something many users may not want). Or people upset about getting AI summaries in search that they didn't ask for. And yet, swaths of capital has been invested in these ideas and perhaps its a worthwhile use of resources. I am not sure personally. Time will tell. But I suspect its more complicated than the author is implying here.
Perhaps we should judge the performance of an LLM by how well it can compress arbitrary information. A higher IQ would mean more compression, after all.
Which LLMs perform better or worse will be determined entirely by the scoring formula used and how it penalizes errors. It is not in the nature of an LLM to be capable of lossless compression.
Was it? I don't remember ever running into anyone preferring long documents. Also, anything added by the LLM is pure noise with the possibility of a hallucination or two. If, for some reason, you might even add some relevant information at times, and you're not going to start making things up.
And if an LLM is also used at the other endpoint to parse the longer text, that creates a broken telephone. Congrats, your communication channel is now unreliable.
I think the usage of LLMs will push for a societal change in how we communicate.
instead of elongated sentences, we perhaps might start seeing an increase in just communicating through the minimum constructing points of whatever meaning we hope to convey, leaving the presentation work for the LLM on the receiving side
I saw an interesting argument recently that the reason you get this type of verbose language in corporate settings is that English lacks a formal tense. Apparently it's much less common in languages that have one. But in corporate English the verbosity is used as a signal that you took time to produce the text out of respect for the person you're communicating with.
This of course now gets weird with LLMs because I doubt it can last as a signal of respect for very long when it just means you fed some bullet points to ChatGPT.
I’m a native Spanish speaker—all forms of written Spanish are more verbose than English, but the formal form is even more verbose. I remember notifications my school used to send my parents were hilariously wordy by English standards.
> notifications my school used to send my parents were hilariously wordy
There might be something else at play there. Public sector workers are notorious for wooden language.
The example I hate the most is how they always say "number of" before every number like "we'll buy 10 new busses" becomes "we will commence the procurement of a number of 10 new buses".
Some of this has massively fallen away over the last thirty years. I spent a lot of time on Japanese linguistics, and the really formal tense usage was already falling away even in business context by the mid-1990s. I still find it fun to construct sentences that are practically gramnatical self-abasement but it's not common in actual spoken or written Japanese.
The 4 paragraphs requirement was not introduced 'because LLM'. It was there all along for what just should have been 'gimme 2 -3 bullet points'. They wanted Bob to hold back on requesting the new machine he needed, not by denying his request openly, but by making the process convoluted. Now Bob can cut through the BS, they want to blame the LMM for wasting their time and resources? BS!
Maybe, but another way to look at this is that if someone is going to hold back requesting a new machine because they don't feel like writing a few paragraphs, then they probably don't really need a new computer. On the other hand, if their current machine really is so bad that it's getting in the way of their work they won't hesitate to quickly bang out 4 paragraphs to get a new one. Obviously, this trick does not work with LLMs in the mix.
Or perhaps they desperately need a new computer but keep putting it off because they are busy with other tasks. And they'd be slowed down, at significant cost to the company until they get the free time to fill out the form.
Furthermore, the guy who doesn't need a new computer is probably going to be writing the paragraphs on company time, so he'd do it if he has nothing to do and needs to look busy. (I think it's safe to assume that the sort of place where you need 4 paragraphs to justify a new computer would also require for employees to look busy at all times.)
Why is everyone trying so hard to find purpose in broken administrative processes? There's usually none and, if there is its usually so hideous that it can't be put into writing.
>Creating the necessary prose is torturous for most of us, so Bob fires up the LLM du jour, types in “Please create a 4 paragraph long business case for my manager, explaining why I need to replace my old, slow computer” and copies the result into his email.
>Bob’s manager receives 4 paragraphs of dense prose and realises from the first line that he’s going to have to read the whole thing carefully to work out what he’s being asked for and why. Instead, he copies the email into the LLM du jour and types at the start “Please summarise this email for me in one sentence”. The 4 paragraphs are summarised as “The sender needs a new computer as his current one is old and slow and makes him unproductive.”
>something very strange about people writing bullet points, having ChatGPT expand it to a polite email, sending it, and the sender using ChatGPT to condense it into the key bullet points 2:42 PM · Mar 2, 2023 · 1.2M Views
It's not really strange, is it? Business and politeness etiquette requires a certain phrasing that is typically more tedious than the core information of a message.
Now that decorating any message with such fluff is automated, we can as well drop the requirement and just state clearly what we want without fluff.
I do not know about where these people live, but managers did not read long emails for years already. Not that blame them, but this world where they would actually want those 4 paragraphs essays did not existed for years.
Because of this, I tend to structure my business communications as "Sentence or five that best conveys what I want from (or want to tell) the reader, followed by the required background and/or supporting information required to understand the problem being described, or justify the request being made.".
However, you can't do much of anything to deal with people who stop reading at the first sentence or question. Those folks are hopeless.
One of the things that makes me hopeful for the future of LLMs is precisely this: humans are needlessly verbose, and LLMs can cut through the crap.
I expect smaller models to become incrementally better at compressing what truly matters in terms of information. Books, reports, blog posts… all kinds of long-form content can be synthesized in just a few words or pages. It’s no wonder that even small LLMs can provide accurate results for many queries.
Oh, no, I do understand the value of long-form and deep human communication. I’m an avid book reader, for instance, and I actually prefer longer narratives.
What I don’t agree with is being needlessly verbose in circumstances in which the opposite is more valuable. Unfortunately, humans have a tendency to use flowery language even when that comes at the expense of message clarity.
Think, for example, of the countless self-help books that can be converted to short blog posts. Or think about legal or academic writing, which often stands in the way of readers actually understanding what is being said.
There’s so much writing like this out there that even LLMs were notorious for taking this over-elaborate language to a whole new level. And in my opinion that’s the kind of thing that we can (and should) avoid.
People are not reading and absorbing the longer pieces of writing. I worked at a company which would create information-dense, wordy PowerPoint slides for client updates, and at some point you see it has no point besides trying to impress. If we actually needed something in an email, they told us to put it in the first sentence.
I've also noticed it when people post LLM writing on Reddit. Something may give me pause, and then re-reading the content any given paragraph was way off. I had even glossed over the bolded conclusion "it’s a coaching-wheels moment" (?) because as you read it your brain thinks of a way it could make sense.
The compression-accuracy tradeoff is fundamental - as models get better at summarizing, they inevitably discard information deemed "less relevant," which introduces risk when the discarded details actually matter in specific contexts.
Why choose the word "inflation" to mean the opposite of compression? If you said it to a stranger, they'd assume you mean the price of LLMs is going up due to scarcity. I would call this LLM fluffing or LLM decompression
My former (obviously) wannabe manager used GAI to pimp our CV's before sending out to clients, pretty sure they too consulted stupid to summarize on their end.
The problem described in this post has nothing to do with LLMs. It has everything to do with work culture and bureaucracy. Rules and laws that don't make sense remain because changing it requires time, energy and effort that most people in companies have either tried and failed or don't care enough to make a change.
This is one example of the "horseless carriage" AI solutions. I've begun questioning further that actually we're going into a generation where a lot of the things we are doing now are not even necessary.
I'll give you one more example. The whole "Office" stack of ["Word", "Excel", "Powerpoint"] can also go away. But we still use it because change is hard.
Answer me this question. In the near future if we could have LLMs that can traverse to massive amount of data why do we need to make excel sheets anymore? Will we as a society continue to make excel spreadsheets because we want the insights the sheet provides or do we make excel sheets to make excel sheets.
The current generation of LLM products I find are horseless carriages. Why would you need agents to make spreadsheets when you should just be able to ask the agent to give you answers you are looking for from the spreadsheet.
> Answer me this question. In the near future if we could have LLMs that can traverse to massive amount of data why do we need to make excel sheets anymore?
A couple of related questions- if airplanes can fly themselves with auto-pilot, why do we need steering yolks? If I have a dishwasher- why do I still keep sponges and dish soap next to my sink?
The technology is nowhere near being reliable enough that we can eschew traditional means of interacting with data. That doesn't prevent the technology from being massively useful.
Think of them as an artifact of a snapshot of time. You can sign them and file them away, perform backups on them, and use that document the intent at that time.
LLMs are not able to replace Excel in their current state. See this simple accounting test: https://accounting.penrose.com/ - errors compound over time. This is the case even with small datasets. For massive corporate datasets it's useless (try asking Gemini to summarise a Google Sheet).
Until there is a fix for this (not clear there ever will be), Excel will be necessary.
Word will probably become a different, more collaborative product. Notion-esque.
Powerpoint...I would love if it disappeared but ultimately if you have to present something, you need to have done the work.
> Why would you need agents to make spreadsheets when you should just be able to ask the agent to give you answers you are looking for from the spreadsheet.
Because it seems to be a fundamental property of LLMs that they just make things up all the time. It's better to make the LLM a natural interface to a formal query language which will return hard answers with fidelity from the database.
> At best we’re implicitly rewarding obfuscation and time wasting; at worst we’re allowing a lack of clear thinking to be covered up.
Most people don't think very clearly. That's why rhetoric is effective. That's why most communication is fluffy social signaling. You can give people great advice and their eyes glaze over because the words didn't fill them with emotion, or something, and they do the exact opposite.
No wonder LLMs get put to work playing that stupid game.
> Bob needs a new computer for his job. In order to obtain a new work computer he has to create a 4 paragraph business case explaining why the new computer will improve his productivity.
Is this situation in any way realistic one? Because the way companies work in my beck of woods, no one wants your 4 paragraph business case essay about computer. Like, it is funny anecdote.
But, in real world, at least in my experience, pretty much everyone preferred short for emails and messages. They would skim the long ones at best, especially in situation that can be boiled down to "Tom wants a new computer and is verbose about it".
You give the concise version to the person who is going to authorise your request. The four paragraph version goes on record for the people that person needs to justfy the descision to, they'll likely declare “I don't see a problem here” without actually reading it which is the intention: they might be more wont to question the shorter version.
My point is, I never encountered this. Literally never. I am not particularly young and it is not even the case that I would work in startups all that much. My friends did not complained about having to write paragraphs essays for stuff like this either.
I am open to the idea that there is some bureaucratic workplace where it works like that ... but everywhere I have experience with, they preferred the short version.
No, it's much worse than that. In real life you talk about pages and pages of documents and power points and meetings after meetings if you happen to need a computer/server/configuration that's not in the pre-approved list. (I really wish I was exaggerating. And of course no, not all employers are like this to state the obligatory obvious.)
> Bob needs a new computer for his job.... In order to obtain a new work computer he has to create a 4 paragraph business case explaining why the new computer will improve his productivity.
> Bob’s manager receives 4 paragraphs of dense prose and realises from the first line that he’s going to have to read the whole thing carefully to work out what he’s being asked for and why. Instead, he copies the email into the LLM.... The 4 paragraphs are summarised as “The sender needs a new computer as his current one is old and slow and makes him unproductive.” The manager approves the request.
"LLM inflation" as a "bad" thing often reflects a "bad" system.
In the case described, the bad system is the expectation that one has to write, or is more likely to obtain a favorable result from writing, a 4 paragraph business case. Since Bob inflates his words to fill 4 paragraphs and the manager deflates them to summarise, it's clear that the 4 paragraph expectation/incentive is the "bad" thing here.
This phenomenon of assigning the cause of "bad" things to LLMs is pretty rife.
In fact, one could say that the LLM is optimizing given the system requirement: it's a lot easier to get around this bad framework.
The 4-paragraph business case was useful for creating friction, which meant that if you couldn't be bothered to write 4 paragraphs you very likely didn't need the computer upgrade in the first place.
This might have been a genuinely useful system, something which broke down with the existence of LLMs.
The only definitively non-renewable resource is time. Time is often spent like a currency, whose monetary instrument is some tangible proxy of how much time elapsed. Verbosity was an excellent proxy, at least prior to the advent of generative AI. As you said, the reason Bob needs to write 4 paragraphs to get a new PC is to prove that he spent the requisite time for that computer, and is thus serious about the request. It’s the same reason management consultants and investment bankers spend 80+ hours a week working on enormous slide decks that only ever get skimmed by their clients: it proves to the clients that the firm spent time on them, and is thus serious about the case/deal. It’s also the same reason a concise thank-you note “thanks for the invite! we had a blast!” or a concise condolence note “very sorry for your loss” get a lot less well-received than a couple verbose paragraphs on how great the event was or how much the deceased will be missed, even if all that extra verbiage confers absolutely nothing beyond the core sentiment. (The very best notes, of course, use their extra words to convey something personally meaningful beyond “thanks” or “sorry.”)
Gen-AI completely negates meaningless verbosity as a proxy of time spent. It will be interesting to see what emerges as a new proxy, since time-as-currency is extremely engrained into the fabric of human social interactions.
There's an important asymmetry here: it takes a lot of to weave an intricate pattern, but much less time to assess and even appreciate it. The sender / suitor pays significantly more than the receiver / decider.
There are some parallels to that in compression and cryptography, but they are rather far-fetched.
A proof of time spent cryptocurrency is the only way I can see a way out of this.
I think requiring in-person physical interactions is more likely
This is the sort of workplace philosophising that I hate the most. Employees aren't children. They don't need to have artificial bullshit put up in between them and what they need, the person approving just needs to actually pay attention.
If someone wants a new computer they should just have to say why. And if it's a good reason, give it to them. If not, don't. Managers have to manage. They have to do their jobs. I'm a manager and I do my job by listening to the people I manage. I don't put them through humiliation rituals to get new equipment.
Sounds like you did not work all that much.
People want new computer because new hire Peter got a new one.
People want new computer because they just had lunch with a friend that works in a different company and got a new computer and they just need to one up him next time they go to lunch.
That is why I am not going to just give people computers because they ask. Worst crybabies come back because they „spilled coffee” on perfectly fine 2 years old laptop.
Wooo I used to think this was how managers work and just ... Was inevitable. I'm so glad to actually be a manager now, because no, it's not. You don't call people who spill coffee, (have you never spilled anything?) crybabies. This is a bad manager.
I don’t call them like that it is just a goof to make comment more interesting.
Not everything you read on the internet is 1 to 1 real life ;)
reread the comment you're responding to because I don't think you read it all the way. Specifically this part:
>If someone wants a new computer they should just have to say why. And if it's a good reason, give it to them. If not, don't.
People make up reasons, no one comes in saying „Peter has a new laptop I want one as well”.
4 paragraph essay wont solve issue with Peter having better computer. Competitive person will write it and make up a reasons.
If you are awarding computers based on 3 paragraph essays, you are having horribly inefficient process that rewards creative writing rather then work.
Which is why none of it happen in real companies, unless they are someones startup expected to fail anyway.
I can say when I wanted to move my desk from one place to another I had to write up the "business justification" (I was already working from both offices on different days and still am, it was a change on paper)
so I'm sure there's large corps that do this for everything. probably ones where you're not asking your manager, but asking finance or IT for everything
The problem is, I'm a verbose writer and can trivially churn out 4 paragraphs - another person is going to struggle. The friction is targeting the wrong area: this is a 15 minute break for me, and an hour long nightmare for my dyslexic co-worker.
Social media will give you a good idea what sort of person enjoys writing 4 paragraphs when something goes wrong; do you really want to incentivize that?
I mean it’s all up to the employer if they want employees to be productive.
If they don’t care they don’t care. They pay most of us for our time anyway, not what we achieve.
I love this article for how it gets the thinking, and I love your response.
I've been aware of similar dynamic in politics, where the collective action/intelligence of the internet destroyed all the old signals politicians used to rely on. Emails don't mean anything like letters used to mean. Even phone calls are automated now. Your words and experience matter more in a statistical big data sense, rather than individually.
---
This puts me in sci-fi world-building mode, wondering what the absurd extension is... maybe it's just proving burned time investment. So maybe in an imagined world where LLMs are available to all as extensions of thought via neural implant, you can't be taken seriously for even the simplest direct statements unless you prove your mind sat and did nothing (aka wasted it's time) for some arbitrary period of time. So if you sat in the corner and registered inactive boredom for 2h, and attached a non-renewable proof of that to a written word, then people would take your perspective seriously, because you expended (though not "gave") your limited attention/time to the request for some significant amount of time
Politics is the OG bad system.
Because politicians literally write the rules of the system, it's incredibly difficult to prevent abuse, bad incentives, and inefficiency.
One of the most fundamentally broken aspects of the US system is that:
1. politicians stay in power by being elected.
2. people are not required to vote and casting a vote incurs a cost (time, travel, expense), therefore not everyone votes.
3. politicians just have to get the votes of the fraction of people who actually do vote and can ignore those who don't.
Politicians are just an extension of the ruling class. Under capitalism, that is large capitalists.
What may seem like abuse and bad incentives to YOU are perfectly aligned goals for THEM.
There's a 1992 modern art piece that was a blank canvas that the artist promised they spent 1000 hours staring directly at
https://www.mirandawhall.space/1000-hours-of-staring/
The broken thing here is that Bob, costing $10k a week, is after a new computer costing $100 a week.
This part I always found funny. Significantly increase productivity of the team for a fraction of the price of employing them? Absolutely not.
I'm a consultant, and have been for 20 years now (except for a 2 year stint as an employee for the benefits and relocation assistance to move my family from Texas to NYC).
At the NYC employment job, I was on a 2 year upgrade cycle being a software developer. Just whatever the current year Dell corpo laptop was. No procurement procedures, just IT got 2 year replacements, our laptops went to the non IT workforce after reimaging.
As a consultant I usually bring my own device, and my laptops are usually WAY more capable since I will run each client on their own VM - makes it easy to delete the client when the contract is up. But I've had one client who did not allow BYOD, and any billed work had to be done on their hardware. That was fine, except that the desktop I was given was already a 12 year old dual core non-hyperthreading CPU that wasn't meant for developers even when it was built. I begged and pleaded for 6+ months for me to either bring in my own hardware they could image now and wipe at the end of the contract, or to please buy me a PC from this decade.
It took 3 years to get the budget approval for a $2000 tower, roughly the equivalent of 15 hours of pay. The thing that finally pushed it over the edge, was that my PC could not handle Teams + Visual Studio at the same time, and manager couldn't handle that he couldn't watch me program.
All of that to say I doubt these non-data-driven organizations are basing these decisions on anything other than micromanagement. Nothing to do with measured or assumed productivity, nothing to do with costs, so all I can think is they have to be a "decision maker" on all aspects.
I never understood why no company I worked at could hire someone just to manage the kitchen instead of paying engineers massive salaries to argue about unloading the dishwasher.
If Bob's job is anything like mine, Bob's new computer will take a week to set up
Sounds like your company needs to spend some time working on their onboarding process. Probably more effective than just hoping no one ever needs a new computer.
this is the biggest deterrent to me requesting a new computer at work.
I used to get hype for laptop refresh days, but I'll ignore them now if my stuffs working fine just to spare myself the nuisance.
Whether it's an additional $100/week for base salary or $100/week for overhead Bob is always going to be after another $100/week regardless what his current cost is. If Bob wants to manage the overheads directly (most don't) like salary costs then he probably wants more of a contractor type position.
As to the content of the letter, the 4 paragraphs are supposed to be "these are reasons I think were missed and why it'll cost more to not correct it" not just "I put effort to write 4 paragraphs of stuff" friction alone.
Having run a short stint as an internal IT manager at an IT focused company... it's astounding how many non-standard/out-of-cycle laptop request are actually either basic user error (even for the most brilliant technical employees) or basic IT systems problems (e.g. poorly tested management/security tool changes eating up performance/battery in certain configurations) that a new laptop won't actually solve. E.g. reports of "my battery runs out in 2 hours and my IM is dog slow" but they are on an M1/M2 MacBook Pro and probably wouldn't notice if they got an M1 or M4 MacBook back as their issue isn't actually the hardware. When someone writes an email or ticket explaining why their use case just wasn't accounted for it's generally pretty obvious they really do need something different.
> Bob’s manager receives 4 paragraphs of dense prose and realises from the first line that he’s going to have to read the whole thing carefully to work out what he’s being asked for and why. Instead, he copies the email into the LLM.... The 4 paragraphs are summarised as “The sender needs a new computer as his current one is old and slow and makes him unproductive.” The manager approves the request.
Bob’s manager is lazy and or an idiot.
Probably both.
> In fact, one could say that the LLM is optimizing given the system requirement: it's a lot easier to get around this bad framework.
Sure, as long as we completely disregard the water, power and silicon wasted to accomplish this goal.
Engaging with why we might actually want inflation of text:
1) For pedagogical or explanatory purposes. For example, if I were to write:
> ∀x∈R,x^2≥0
I've used 10 characters to say
> For every real number x, it's square is greater than or equal to zero
For a mathematician, the first is sufficient. For someone learning, the second might be better (and perhaps as expansion of 'real number' or that 'square' is 'multiplying it by itself').
2) To make sure everything is stated and explicit. "He finally did x" implies that something has been anticipated/worked on for awhile, but "after a period of anticipation he did x" makes it more clear. This also raises the question of who was anticipating, which could be made explicit too.
As someone who spends a lot of time converting specifications to code (and explaining technical problems to non-technical people), unstated assumptions are very prevalent. And then sometimes people have different conceptions of the unstated assumption (i.e. some people might think that nobody was anticipating, it just took longer than you'd expect otherwise).
So longer text might seem like a simple expansion, but then it ends up adding detail.
I definitely agree with the authors point, I just want to argue that having a text-expander tool isn't quite as useless as 'generate garbage for me'.
Can a generator do things like 2 if it wasn't in the input text?
Ambiguity is resolved via provided context, but just as with conversations, this context may be severely underspecified.
Yes, because generators generate at the token level, which is technically smaller than an individual word. They can easily generate unique sentences, and for example transfer learning allows them to apply knowledge obtained from some other training data to new domains.
The idea that generators are some sort of parrot is very outdated. The 2021 paper that coined the term "stochastic parrot" was already wrong when it was published.
> Yes, because generators generate at the token level, which is technically smaller than an individual word. They can easily generate unique sentences, and for example transfer learning allows them to apply knowledge obtained from some other training data to new domains.
Sure. But can they read the original author's mind, and therefore generate the right unique sentence that expresses the actual intent?
Sure, but in the case of "he finally did X" without context passed in, how does the llm determine if it would be expanded as "this was a very anticipated change" or if the author is frustrated at how long it took? If the nuanced meaning isn't there in the input context.
obviously it can generate the longer message, but is it going to go look up what the sentence refers to and infer extra meaning automatically...?
If I need it expanded I can put it into my LLM myself.
This image originally came out just around the time of ChatGPT release and captures it well: https://i.imgur.com/RHGD9Tk.png
This one is my favorite: https://marketoonist.com/2023/03/ai-written-ai-read.html
The important part is the GDP is now increased because of the cost of energy and additional hardware needed expand and then compress the original data. Think of the economic growth all these new hassles provide!
Next innovation: compress the AI translation layer from both sides. I feel like there might be an unbelievable Weissman score that can be achieved!
grug brain take
The older I get the more concise I find myself (which is not to say I'm actually concise, as my comment history will demonstrate), but LLM's have really driven home just how much noise day to day communication involves. So much filler text.
It still surprises me when I see non-technical enthusiasts get excited about LLMs drafting almost useless copy or email or whatever. So much garbage text no one reads but has to be written for some reason. Its weird.
"I wrote this mail slightly longer because I didn't have time to make it short" - someone famous
When writing something I want people to read, I always take time at the end to make it shorter - remove distracting sentences, unnecessary adjectives and other noise. Really works wonders for team communication.
>When writing something I want people to read, I always take time at the end to make it shorter - remove distracting sentences, unnecessary adjectives and other noise.
This is a good advice. How can I do it when talking? I often talk too much saying little, often loosing Listener's attention in the process.
Just like weight loss, start talking less.
Silence is better than useless noise.
On one side you have people using LLMs to fluff a sentence in to an essay. And on the receiver side they are hitting a button to AI summarise it back to a sentence.
What incredible technology.
An LLM is effectively a compressed model of its input data.
Inference is then the decompression stage where it generates text from the input prompt and the compressed model.
Now that compressing and decompressing texts is trivial with LLMs, we humans should focus - in business at least - on communicating only the core of what we want to say.
If the argument to get a new keyboard is: "i like it", then this should suffice, for inflated versions of this argument can be trivially generated.
What I hate about this is that often a novel and interesting idea truly needs extra space to define and illustrate itself, and by virtue of its novelty LLMs will have substantially more difficulty summarizing it correctly. But it sounds like we are heading to a medium-term where people cynically assume any long email must be LLM-generated fluff, and hence nothing is lost by asking for an LLM summary.
What a horrible technology.
Not to be overly snide, but I can imagine that almost every person who writes long, tedious emails that wax on and on thinks they have something novel and interesting that truly needs extra space. Also, most novel things are composed of pedestrian things, which LLMs have no issue summarizing sufficiently.
Maybe you can provide an example where this case would occur, and maybe some indication how often you think this would occur.
It seems to be a big problem for AI-generated summaries of scientific research, where the top-performing AIs ignores key details about 25% of the time, and often much worse. The problem is exactly what I said: the research has a new idea which is (by necessity) not going to be in the pretraining data, so the AI ignores or misstates it. I don't expect this problem to be solved anytime soon, it is inherent to how 2020s artificial neural networks work.
https://royalsocietypublishing.org/doi/10.1098/rsos.241776
Edit: I was thinking about the “overly snide” and am reminded of Sam Bankman-Fried:
The problem is not snideness, it is arrogant cynicism leading to stupidity.> If the argument to get a new keyboard is: "i like it", then this should suffice
This seems like exactly what LLMs are supposed to be good at, according to you, so why don't they just near-losslessly compress the data first, and then train on that?
Also, if they're so good at this, then why are their answers often long-winded and require so much skimming to get what I want?
I'm skeptical LLMs are accurately described as "near lossless de/compression engines".
If you change the temperature settings, they can get quite creative.
They are their algorithm, run on their inputs, which can be roughly described as a form of compression, but it's unlike the main forms of compression we think of - and it at least appears to have emergent decompression properties we aren't used to.
If you up the lossy-ness on a JPEG, you don't really end up with creative outputs. Maybe you do by coincidence, and maybe you only do with LLMs - but at much higher rates.
Whatever is happening does not seem to be what I think people typically associate with simple de/compression.
Theoretically, you can train an LLM on all of Physics, except a few things, and it could discover the missing pieces through reasoning.
Yeah, maybe a JPEG could, too, but the odds of that seem astronomically lower.
If you don't design your compressor to output data that can be compressed further, it's going to trash compressibility.
And if you find a way to compress text that isn't insanely computationally expensive, and still makes the compressed text compressible by LLMs further - i.e. usable in training/inference? You, basically, would have invented a better tokenizer.
A lot of people in the industry are itching for a better tokenizer, so feel free to try.
The inverse of this is "AI Loopidity" where we burn cycles inflating then deflating information (in emails, say, or in AI code that blows up then gets reduced or summarized). This often also leads to weird comms outcomes, like saving a jpg at 85% a dozen times.
And each cycle introducing error like a game of telephone
And with no governance in place (like an engine governor) you get cascade failure.
Be more trivial
I consider inflation a double insult. (https://ruudvanasseldonk.com/2025/llm-interactions) It says "I couldn't be bothered to spend time writing this myself, but I'm expecting you to read all the fluff."
To be fair, the recipient probably isn't reading it either. They're getting it summarized. LLMs are creating more work for other LLMs.
> That we are using LLMs for inflation should not be taken as a criticism of these wonderful tools. It might, however, make us consider why we find ourselves inflating content. At best we’re implicitly rewarding obfuscation and time wasting; at worst we’re allowing a lack of clear thinking to be covered up. I think we’ve all known this to be true, but LLMs allow us to see the full extent of this with our own eyes. Perhaps it will encourage us to change!
Yeah, this is the problem. Wealth distribution stopped working sometime in the late 20th century and we're fighting each others for competitive advantages. That's the core of this phenomenon.
No one needs containers full of baby sized left shoes, but proof of work must be shown. So the leathers must be cut and shoes must be sewn, only to be left in the ever growing pile in the backyard. That's kind of wrong.
Long documents in business contexts that get summarized and go mostly unread are the byproduct of a specific and common level of trust and accountability in those contexts: people don't believe someone has done enough critical thinking or has a strong enough justification for a proposal unless they've put it on the page, but if it is on the page, it's assumed that it does in fact represent critical thinking and legitimate justification.
If trust was higher, shorter documents would be more desirable. If trust was lower, or accountability higher, summarization would be used a lot more carefully.
LLMs haven't changed anything in this regard except that they've made it extremely easy to abuse trust at that specific level. The long-term result will be that trust will fall in the general case, and people will eventually become more careful about using summarization. I don't think it will be long before productized AI used in business contexts will be pretrained/fine-tuned to perform a basic level of AI content detection or include a qualitative measure of information density by default when performing summarization.
Huh, this post is not what I thought it would be! Even after the first two paragraphs!
There's a line of thought which states that intelligence rhymes with compression: Identifying patterns allows better prediction, enables better compression of the data.
However, internally, LLMs typically do the opposite: Tokenization and vectorization multiply the bit rate of the input signal. Chain of thought techniques add a lot of extra text, further increasing the bit rate.
My PM said they’d written a bunch of tickets for a project yesterday morning that we hadn’t fully scoped yet. I was pleasantly surprised because I can’t complain if they are going to get ahead of things and start scaffolding tickets.
Of course when I went to read them they were 100% slop. The funniest requirement were progress bars for actions that don’t have progress. The tickets were, even if you assume the requirements weren’t slop, at least 15 points a piece.
But ok maybe with all of these new tools we can respond by implementing these insane requirements. The real problem is what this article is discussing. Each ticket was also 500-700 words. Requirements that boil down to a single if statement were described in prose. While this is hilarious the problem is it makes them harder to understand.
I tried to explain this and they just said “ok fine rewrite them then”. Which I did in maybe 15min because there wasn’t actually much to write.
At this point I’m at a loss for how to even work with people that are so convinced these things will save time because they look at the volume of the output.
Ask an llm for a project plan and they’ll happily throw dates around for each step, when they can’t possibly know how long it will take.
But project plan dates have always been fiction. Getting there faster is an efficiency win.
That said I’ve found that llms are good as interrogators. If used to guide a conversation, research background information and then be explicitly told to tersely outline the steps in something I’ve had very good results.
The date/week estimations in plans are especially funny when you work with an agent and it spits that out. "Week 1, setting up the structure" - uh, no, we're going to do this in 10 minutes.
> At this point I’m at a loss for how to even work with people that are so convinced these things will save time because they look at the volume of the output.
The same way, presumably, that one used to work with people who would say things like "just look how much code this template system generates for us!" unironically.
The software requirements phase is becoming increasingly critical to the development lifecycle and that trend will continue. I have started writing very short tickets and having claude code inflate them, then I polish those. I often include negative prompts at this point so claude may have included "add a progress bar for xyz" and i simply add "do not" in front of those things that do not make sense. The results have been excellent.
It's disheartening to hear that people get paid six figures to implement things like "progress bars."
Why? As a frequent consumer of software I love progress bars and I know many others do too. Are you mad other people get paid highly or mad about progress bars? or some combination of both?
If progress bars are actually showing information I want, like how much time is left on the operation, they are great! I suspect most are just garbage though.
The only acceptable response to obvious AI slop - unless it's it's clear it's been heavily reviewed and updated - is to put it back into the AI and ask it for a 1 paragraph summary and work off of that.
> the load on my server is reduced
isn't this the opposite? Enabling compression will INCREASE the load on your server as you need more CPU to compress/decompress the data.
As Koffiepoeder suggests, since the vast majority of content on my site is static, I only have to compress a file once when I build the site, no matter how many people later download it. [The small amount of dynamic content on my site isn't compressed, for the reason you suggest.]
That’s a good point, didn’t know it was cached on top.
Depends on efficiency of compression I guess. If X bytes of data takes N time to transmit, and each slice of N takes Y CPU cycles during transmission, how many Y must your compression algorithm use, and how low must it lower N, in order to be more efficient from a CPU utilisation perspective? assumedly there's an inflection point from CPU use perspective, maybe a point that's impractical to achieve? I'm just vibe-thinking-out-loud.
Or your server can cache the compressed content (since it is a static page anyway).
It depends on where the bottleneck is. If it’s in network packet size, it would help serve more clients. At the expense of more CPU needed to decode/encode the data. If you’re serving large files and you have headroom in hardware it’s totally normal.
Not necessarily. For example you can pre-compress your files once, and then it'll be up to the clients to decompress on receipt.
LLM Inflation is an interesting choice of terminology. Like with many things in our contemporary society there is a temptation to assign absolute, quantitative value to everyday concepts but realistically we should know this to be a fallacy. Many concepts actually have no "on paper" value but still manifest significant social value. Marketing is the typical example, and yet we don't refer to advertising as inflation (though maybe we should).
This concept probably applies to lots of work in the "AI" space right now. The idea of using huge amounts of compute to generate lifelike voices for LLMs comes to mind as being recently maligned (something many users may not want). Or people upset about getting AI summaries in search that they didn't ask for. And yet, swaths of capital has been invested in these ideas and perhaps its a worthwhile use of resources. I am not sure personally. Time will tell. But I suspect its more complicated than the author is implying here.
Perhaps we should judge the performance of an LLM by how well it can compress arbitrary information. A higher IQ would mean more compression, after all.
Which LLMs perform better or worse will be determined entirely by the scoring formula used and how it penalizes errors. It is not in the nature of an LLM to be capable of lossless compression.
We already do this.
Prediction is formally equivalent to compression, so loss is just a measure of how well you can compress the training dataset.
In a sense, yes. But that's compression of an input/output map, not arbitrary information. Also, it is not lossless.
Lossily or losslessly?
Losslessly, because that's easier to test.
When you're recalling a memory, do you remember the position of every blade of grass, or the exact angle of the Sun?
Humans, the only extant example of a general intelligence, don't do lossless compression at all.
I don't think you get to AGI by trying to compress noise.
Lossless compression requires the recognition of patterns in the data. And a smart way to use those patterns to perform the actual compression.
This type of verbiage inflation was happening in business all the time anyway. LLMs are just being used as a method for doing it faster.
Was it? I don't remember ever running into anyone preferring long documents. Also, anything added by the LLM is pure noise with the possibility of a hallucination or two. If, for some reason, you might even add some relevant information at times, and you're not going to start making things up.
And if an LLM is also used at the other endpoint to parse the longer text, that creates a broken telephone. Congrats, your communication channel is now unreliable.
We've also had XML inflation for 30 years.
I think the usage of LLMs will push for a societal change in how we communicate.
instead of elongated sentences, we perhaps might start seeing an increase in just communicating through the minimum constructing points of whatever meaning we hope to convey, leaving the presentation work for the LLM on the receiving side
I saw an interesting argument recently that the reason you get this type of verbose language in corporate settings is that English lacks a formal tense. Apparently it's much less common in languages that have one. But in corporate English the verbosity is used as a signal that you took time to produce the text out of respect for the person you're communicating with.
This of course now gets weird with LLMs because I doubt it can last as a signal of respect for very long when it just means you fed some bullet points to ChatGPT.
Seems like an easy hypothesis to test: Do languages with a formal tense have short corporate language?
I’m a native Spanish speaker—all forms of written Spanish are more verbose than English, but the formal form is even more verbose. I remember notifications my school used to send my parents were hilariously wordy by English standards.
> notifications my school used to send my parents were hilariously wordy
There might be something else at play there. Public sector workers are notorious for wooden language.
The example I hate the most is how they always say "number of" before every number like "we'll buy 10 new busses" becomes "we will commence the procurement of a number of 10 new buses".
Public sector workers? No, this was a private school.
I can speak 4 EU languages besides english. All 4 have special forms which are “formal” all 4 more verbose in the formal form. So if you ask me: “no”
This is what the argument I read claimed, I haven't verified it.
Japanese has polite forms, but business emails are anything but shorter than in English: https://shiftasia.com/community/choose-the-right-email-greet...
Some of this has massively fallen away over the last thirty years. I spent a lot of time on Japanese linguistics, and the really formal tense usage was already falling away even in business context by the mid-1990s. I still find it fun to construct sentences that are practically gramnatical self-abasement but it's not common in actual spoken or written Japanese.
The example in the article does not look like LLM Inflation, but that LLM can't reduce the waste in a bureaucratic process.
I call BS.
The 4 paragraphs requirement was not introduced 'because LLM'. It was there all along for what just should have been 'gimme 2 -3 bullet points'. They wanted Bob to hold back on requesting the new machine he needed, not by denying his request openly, but by making the process convoluted. Now Bob can cut through the BS, they want to blame the LMM for wasting their time and resources? BS!
Maybe, but another way to look at this is that if someone is going to hold back requesting a new machine because they don't feel like writing a few paragraphs, then they probably don't really need a new computer. On the other hand, if their current machine really is so bad that it's getting in the way of their work they won't hesitate to quickly bang out 4 paragraphs to get a new one. Obviously, this trick does not work with LLMs in the mix.
Or perhaps they desperately need a new computer but keep putting it off because they are busy with other tasks. And they'd be slowed down, at significant cost to the company until they get the free time to fill out the form.
Furthermore, the guy who doesn't need a new computer is probably going to be writing the paragraphs on company time, so he'd do it if he has nothing to do and needs to look busy. (I think it's safe to assume that the sort of place where you need 4 paragraphs to justify a new computer would also require for employees to look busy at all times.)
Why is everyone trying so hard to find purpose in broken administrative processes? There's usually none and, if there is its usually so hideous that it can't be put into writing.
>Creating the necessary prose is torturous for most of us, so Bob fires up the LLM du jour, types in “Please create a 4 paragraph long business case for my manager, explaining why I need to replace my old, slow computer” and copies the result into his email.
>Bob’s manager receives 4 paragraphs of dense prose and realises from the first line that he’s going to have to read the whole thing carefully to work out what he’s being asked for and why. Instead, he copies the email into the LLM du jour and types at the start “Please summarise this email for me in one sentence”. The 4 paragraphs are summarised as “The sender needs a new computer as his current one is old and slow and makes him unproductive.”
Sam Altman actually had a concise tweet about this blog's topic (https://x.com/sama/status/1631394688384270336)
>something very strange about people writing bullet points, having ChatGPT expand it to a polite email, sending it, and the sender using ChatGPT to condense it into the key bullet points 2:42 PM · Mar 2, 2023 · 1.2M Views
It's not really strange, is it? Business and politeness etiquette requires a certain phrasing that is typically more tedious than the core information of a message.
Now that decorating any message with such fluff is automated, we can as well drop the requirement and just state clearly what we want without fluff.
I do not know about where these people live, but managers did not read long emails for years already. Not that blame them, but this world where they would actually want those 4 paragraphs essays did not existed for years.
Because of this, I tend to structure my business communications as "Sentence or five that best conveys what I want from (or want to tell) the reader, followed by the required background and/or supporting information required to understand the problem being described, or justify the request being made.".
However, you can't do much of anything to deal with people who stop reading at the first sentence or question. Those folks are hopeless.
One of the things that makes me hopeful for the future of LLMs is precisely this: humans are needlessly verbose, and LLMs can cut through the crap.
I expect smaller models to become incrementally better at compressing what truly matters in terms of information. Books, reports, blog posts… all kinds of long-form content can be synthesized in just a few words or pages. It’s no wonder that even small LLMs can provide accurate results for many queries.
> humans are needlessly verbose
What a depressing belief. Human communication is about a whole lot more than just getting your point across as quickly and efficiently as possible.
Oh, no, I do understand the value of long-form and deep human communication. I’m an avid book reader, for instance, and I actually prefer longer narratives.
What I don’t agree with is being needlessly verbose in circumstances in which the opposite is more valuable. Unfortunately, humans have a tendency to use flowery language even when that comes at the expense of message clarity.
Think, for example, of the countless self-help books that can be converted to short blog posts. Or think about legal or academic writing, which often stands in the way of readers actually understanding what is being said.
There’s so much writing like this out there that even LLMs were notorious for taking this over-elaborate language to a whole new level. And in my opinion that’s the kind of thing that we can (and should) avoid.
You mean that language is a form or bureaucracy and gate keeping?
People are not reading and absorbing the longer pieces of writing. I worked at a company which would create information-dense, wordy PowerPoint slides for client updates, and at some point you see it has no point besides trying to impress. If we actually needed something in an email, they told us to put it in the first sentence.
I've also noticed it when people post LLM writing on Reddit. Something may give me pause, and then re-reading the content any given paragraph was way off. I had even glossed over the bolded conclusion "it’s a coaching-wheels moment" (?) because as you read it your brain thinks of a way it could make sense.
Pretty often it should be about getting your point across as efficiently as possibly though, but people add fluff out of tradition or cultural norms.
There are many, many, contexts where humans just pile text up for no real reason than to fill space.
The compression-accuracy tradeoff is fundamental - as models get better at summarizing, they inevitably discard information deemed "less relevant," which introduces risk when the discarded details actually matter in specific contexts.
Why choose the word "inflation" to mean the opposite of compression? If you said it to a stranger, they'd assume you mean the price of LLMs is going up due to scarcity. I would call this LLM fluffing or LLM decompression
> Why choose the word "inflation" to mean the opposite of compression?
https://en.wikipedia.org/wiki/Deflate seems relevant here.
Plus there's already a name for the general phenomenon: AI slop.
Sure, it doesn't directly capture the compression/decompression aspect, but it's assumed that slop includes unnecessary filler.
I've seen this happen IRL.
My former (obviously) wannabe manager used GAI to pimp our CV's before sending out to clients, pretty sure they too consulted stupid to summarize on their end.
https://x.com/itaysk/status/1887942925042033069
Where I work, I do the opposite: I let my colleagues know that they should write much mess and much more concise.
I actually straight up reject it when text is too inflated, and I remind people that LLMs are available to expand on request.
> Where I work, I do the opposite: I let my colleagues know that they should write much mess
Sounds like a fun place :-)
(yes, yes, I know it's a typo, I could not resist)
Haha, woops! Well, the thing is that LLMs actually enable being a bit more messy
Good point to better use it for haikus. Not sure if it that good at it though.
The problem described in this post has nothing to do with LLMs. It has everything to do with work culture and bureaucracy. Rules and laws that don't make sense remain because changing it requires time, energy and effort that most people in companies have either tried and failed or don't care enough to make a change.
This is one example of the "horseless carriage" AI solutions. I've begun questioning further that actually we're going into a generation where a lot of the things we are doing now are not even necessary.
I'll give you one more example. The whole "Office" stack of ["Word", "Excel", "Powerpoint"] can also go away. But we still use it because change is hard.
Answer me this question. In the near future if we could have LLMs that can traverse to massive amount of data why do we need to make excel sheets anymore? Will we as a society continue to make excel spreadsheets because we want the insights the sheet provides or do we make excel sheets to make excel sheets.
The current generation of LLM products I find are horseless carriages. Why would you need agents to make spreadsheets when you should just be able to ask the agent to give you answers you are looking for from the spreadsheet.
> Answer me this question. In the near future if we could have LLMs that can traverse to massive amount of data why do we need to make excel sheets anymore?
A couple of related questions- if airplanes can fly themselves with auto-pilot, why do we need steering yolks? If I have a dishwasher- why do I still keep sponges and dish soap next to my sink?
The technology is nowhere near being reliable enough that we can eschew traditional means of interacting with data. That doesn't prevent the technology from being massively useful.
>why do we need to make excel sheets anymore?
Think of them as an artifact of a snapshot of time. You can sign them and file them away, perform backups on them, and use that document the intent at that time.
Audits don't work so well on LLMs
LLMs are not able to replace Excel in their current state. See this simple accounting test: https://accounting.penrose.com/ - errors compound over time. This is the case even with small datasets. For massive corporate datasets it's useless (try asking Gemini to summarise a Google Sheet).
Until there is a fix for this (not clear there ever will be), Excel will be necessary.
Word will probably become a different, more collaborative product. Notion-esque.
Powerpoint...I would love if it disappeared but ultimately if you have to present something, you need to have done the work.
> Why would you need agents to make spreadsheets when you should just be able to ask the agent to give you answers you are looking for from the spreadsheet.
Because it seems to be a fundamental property of LLMs that they just make things up all the time. It's better to make the LLM a natural interface to a formal query language which will return hard answers with fidelity from the database.
> Brevity is the soul of wit
Now that pachinko machines can create lots of prose, maybe it's time to finally learn this lesson.
Bob is highly optimistic!
> At best we’re implicitly rewarding obfuscation and time wasting; at worst we’re allowing a lack of clear thinking to be covered up.
Most people don't think very clearly. That's why rhetoric is effective. That's why most communication is fluffy social signaling. You can give people great advice and their eyes glaze over because the words didn't fill them with emotion, or something, and they do the exact opposite.
No wonder LLMs get put to work playing that stupid game.
> Bob needs a new computer for his job. In order to obtain a new work computer he has to create a 4 paragraph business case explaining why the new computer will improve his productivity.
Is this situation in any way realistic one? Because the way companies work in my beck of woods, no one wants your 4 paragraph business case essay about computer. Like, it is funny anecdote.
But, in real world, at least in my experience, pretty much everyone preferred short for emails and messages. They would skim the long ones at best, especially in situation that can be boiled down to "Tom wants a new computer and is verbose about it".
You give the concise version to the person who is going to authorise your request. The four paragraph version goes on record for the people that person needs to justfy the descision to, they'll likely declare “I don't see a problem here” without actually reading it which is the intention: they might be more wont to question the shorter version.
My point is, I never encountered this. Literally never. I am not particularly young and it is not even the case that I would work in startups all that much. My friends did not complained about having to write paragraphs essays for stuff like this either.
I am open to the idea that there is some bureaucratic workplace where it works like that ... but everywhere I have experience with, they preferred the short version.
> Is this situation in any way realistic one?
No, it's much worse than that. In real life you talk about pages and pages of documents and power points and meetings after meetings if you happen to need a computer/server/configuration that's not in the pre-approved list. (I really wish I was exaggerating. And of course no, not all employers are like this to state the obligatory obvious.)
The author made up a fake situation to drive a point that doesn't exist
> One of the signal achievements of computing is data compression
Ah, yes. It is an achievement in signals in a way.