The tech sector is the most software-defined and virtual of all sectors, it's natural that they are first to change. Other industries need to make capital investments to adapt to change.
You can see a similar history with electrification. A few industries went quickly, but then it took like a decade of little growth before the true benefits showed up. As an example, take a factory that was built around hyrdo-power like a grain mill. You can't just take an electric motor and drop it in in place of the water wheel. It doesn't fit, the wheel is outside, there isn't room between the mill and the wall, etc. Only when you completely rebuild a new mill building designed around the features and limitations of the electric motor do you see the major benefits.
Same thing here; it will be new factories and new companies that are designed from the ground up around AI and its benefits and limitations that really move the needle.
Cool analogy, but there's another possibility: software is uniquely suited for being boosted by LLM-based coding agents. That's not to say LLMs are useless in other domains, but they may not yield as much benefit.
Software creation tools have been worked on for decades to keep fallible human coders out of trouble as much as possible. Starting with compilers that enforce or at least encourage coding style humans can understand. Compilers embody formal definitions of language syntax. Lint and other static analysis technologies contribute by identifying problematic patterns in code. Leak detectors identify memory management problems. Profilers identify performance problems.
Few other endeavors have this much tooling that serendipitously happens to help help LLM-based agents be more useful and less error prone.
A lot of discussion around AI is very high-level and abstract. It helps to dig in and identify the characteristics that actually make a domain more or less amenable to being boosted by AI.
I think LLM's will prove to be useful in any domain where most work happens in front of a computer. As LLM's make programming cheap, they also reduce the cost of building new software that automates work currently performed by humans in other fields.
If you provide best models with all the data and context an average white collar worker has to perform their job, in most cases they will do just fine. But in real world that context is often scattered over multiple systems, or sometimes exists only in people's heads. I believe that's the main bottleneck for adoption now.
You're missing the biggest suggestion from the GP's post: that the key to LLM success is verifiability.
Software has high verifiability through compilers, linters, static analysis, leak/memory corruption detection tools, or the very least, running the program to see if it crashes.
Math (as has been making the headlines recently) has high verifiability through automated theorem provers. Notably some of the non-mechanized math results from LLMs have been demonstrated to be wrong (and presumably many more are being checked; these aren't exactly easy to understand proofs).
Game playing (chess, go, Star Craft, etc.) is similarly verifiable: you get your score and you always know how you did in the end (even if not in the moment).
Making assumptions about LLMs in non-veriable domains is making a claim about LLMs that doesn't match their current capabilities. Maybe they'll get there and maybe they won't; I can speculate as well as you can but we're both guessing, so I'm not sure how much value the speculation has.
A lot of testing and verification was avoided in the past because it wasn’t considered nearly as valuable as shipping code. Many of our jobs are evolving to building verifiers of AI output.
If you believe the hype/marketing around how scary "AI" is LLMs are perfectly happy to go rouge and lie, cheat, and even blackmail to protect themselves.
This assumes that a significant portion of the average white collar worker’s job is amenable to automation.
But most white collar workers I know spend most their time either talking to each other, or thinking real hard about what to say the next time they get together to talk. If that’s generally true then it would suggest that there isn’t much white collar work that LLMs, which are now widely regarded as exceptionally poor communicators, can do particularly well.
I think the bear case (see:short term) benefit is that any and all information gathering, management and parsing the employee did between organizing with other employees will be automated, allowing the employee to do more organizing over the same unit of time.
The bull case (see: long term) is that the “organizing and talking” as long as it’s not something that is essentially human (like caretaking, nursing etc.) will happen between personal agents of sorts with only the deeper judgments being left to the humans opening up even more increases in productivity. This is already happening with PR reviews in a lot of workplaces.
Though I’m guessing most workplaces outside of tech will require total reorganization around this reality, but more likely this way of running will mean newer upstarts challenging incumbents refusing to change as it’s always been.
> thinking real hard about what to say the next time they get together
LLMs can help tremendously with this part, though. Exploration/brainstorming, discovery, data analysis, etc can be made significantly easier with LLMs in most knowledge based fields.
They certainly can, but I wonder if the cost will be deskilling your employees and leading brainstorming sessions towards AI psychosis.
I've seen much time wasted on strange ideas originated by an LLM. It's a lot easier for a human to organically have good ideas than to refute LLM-generated ideas that are a dead end.
If people are in charge of project management, and repeatedly fail to recognise poor ideas and instead pursue them to the detriment of productivity, then they probably shouldn't be managing that project.
All AI is doing here is rapidly identifying the people who have Peter-principled themselves.
> It's a lot easier for a human to organically have good ideas than to refute LLM-generated ideas that are a dead end.
What works better is when you can use an LLM to obtain a negative opinion on your own ideas that you think are good. LLMs aren't great at original thinking, but they're pretty good at finding the flaws in poorly thought-through human ideas.
This is certainly what I saw. I was at a company that was a very eager early adopter of GenAI for all work. One big thing that came out of it was a bunch of effort wasted on unprofitable projects.
The chatbots kept steering people toward uncreative ideas and overshadowing novel insights that might come from genuine domain expertise. AI generated meeting summaries became a force for miscommunication as they de-emphasized or even outright misrepresented what was really important about the conversation. And people became much worse at proving ideas out before making significant resource commitments thanks to effects that we now describe with terms like AI psychosis.
I work for big tech. We don’t have human in the loop anymore. Ai tools handle 99%. It’s not irresponsible if done right. And the next billionaire will show that ai can be tamed, managed, and humans are the problem.
That’s what they say about communism and anarchism and many other isms: if it’s not working, you aren’t doing it right. Every failed implementation gets the No True Scotsman argument.
"Beseech the god machine that it may shower upon you its tokenous beneficence."
The god machine can prompt itself. Just go "Write me a prompt to <whatever>. Ask me as many clarifying questions as needed.". It is not a differentiator.
Deep domain knowledge and pattern matching that allows you to quickly verify outputs might be a differentiator. Maybe. But you might end up the bottleneck if you verify everything.
What differentiation? Most work is bread-and-butter after you get a few years of experience. If you LLM's can automate 50% of that, it would transform society already.
That “after a few years’ experience” bit is the catch. What’s happening during that time is a whole bunch of skill acquisition, building mental domain models, building context- and culture-specific soft skills, etc.
All of it relies on learning and information processing capabilities that LLMs simply do not have. The idea that any of it will automatically emerge from the current architectures simply by adding more parameters is about as sensible as thinking that if you keep counting long enough eventually you’ll get to complex numbers. We’re going to need to invent new technology to fill these gaps, and there is no way of knowing how long those developments will take.
They're in the process of training the models for those professions, too. They're offering people hundreds of dollars an hour to run recordings of workload they create and questions they ask. It's only a matter of time that companies sell access to those models and sell it as that profession.
>But most white collar workers I know spend most their time either talking to each other, or thinking real hard about what to say the next time they get together to talk.
Maybe that's just something which made sense when they were necessary to do the not yet automatable informattion processing work, but that only took like 20% of their time, so they added meetings, and useless work to keep them busy.
I don’t believe there is any fundamental reason LLMs can’t do this task, and the base models are probably already capable of it. RL to elicit the right behavior just hasn’t been done yet. Coding got solved first because it was cheaper and easier to get the necessary data for RL environments, which made it the natural first market opportunity. Solving white collar communication will require more resources to collect the data and build the environments, but the labs now have more resources and are working on it.
I want to believe reasons why this won’t work, but it feels like cope. It would be smart to be prepared for the world where knowledge workers get replaced by AI and also prepared for the world where AI is a bubble that pops.
The goalposts on what neural nets can possibly do have already moved past chess, go, protein folding, coding, poker, math, the Turing test…
"It would be smart to be prepared for the world where knowledge workers get replaced by AI and also prepared for the world where AI is a bubble that pops."
lol aka you have zero conviction about what you are saying.
So whats the point of you posting? Zero value add. In fact bordering on negative value.
You want to drive across town. You, only willing to deal in certainties, must decide ahead of time whether you will get in a crash. If you determine won’t get in a crash, you make the drive, pedal to the metal. If you determine you would crash, you stay home.
Me, willing to deal in probabilities, make the drive at the speed I feel reasonable and wear a seatbelt.
As the evidence from software programming shows, a lot of white collar workers spend most of their time talking frankly because they are inept at execution.
secabeen 11 hours ago
You can see a similar history with electrification. A few industries went quickly, but then it took like a decade of little growth before the true benefits showed up. As an example, take a factory that was built around hyrdo-power like a grain mill. You can't just take an electric motor and drop it in in place of the water wheel. It doesn't fit, the wheel is outside, there isn't room between the mill and the wall, etc. Only when you completely rebuild a new mill building designed around the features and limitations of the electric motor do you see the major benefits.
Same thing here; it will be new factories and new companies that are designed from the ground up around AI and its benefits and limitations that really move the needle.
Zigurd 11 hours ago
Software creation tools have been worked on for decades to keep fallible human coders out of trouble as much as possible. Starting with compilers that enforce or at least encourage coding style humans can understand. Compilers embody formal definitions of language syntax. Lint and other static analysis technologies contribute by identifying problematic patterns in code. Leak detectors identify memory management problems. Profilers identify performance problems.
Few other endeavors have this much tooling that serendipitously happens to help help LLM-based agents be more useful and less error prone.
A lot of discussion around AI is very high-level and abstract. It helps to dig in and identify the characteristics that actually make a domain more or less amenable to being boosted by AI.
rwyinuse 10 hours ago
If you provide best models with all the data and context an average white collar worker has to perform their job, in most cases they will do just fine. But in real world that context is often scattered over multiple systems, or sometimes exists only in people's heads. I believe that's the main bottleneck for adoption now.
eslaught 8 hours ago
Software has high verifiability through compilers, linters, static analysis, leak/memory corruption detection tools, or the very least, running the program to see if it crashes.
Math (as has been making the headlines recently) has high verifiability through automated theorem provers. Notably some of the non-mechanized math results from LLMs have been demonstrated to be wrong (and presumably many more are being checked; these aren't exactly easy to understand proofs).
Game playing (chess, go, Star Craft, etc.) is similarly verifiable: you get your score and you always know how you did in the end (even if not in the moment).
Making assumptions about LLMs in non-veriable domains is making a claim about LLMs that doesn't match their current capabilities. Maybe they'll get there and maybe they won't; I can speculate as well as you can but we're both guessing, so I'm not sure how much value the speculation has.
dwattttt 8 hours ago
And LLMs won't take the easy way out (like so many people) and suppress errors, change tests etc, tweak code until warnings go away?
fuzzzerd 7 hours ago
That's not to say it isn't useful, and if youre on top of it you can guide it away from doing that.
otterley 6 hours ago
dwattttt 5 hours ago
AI is changing how the code is written, but I'm not seeing anything at all changing the incentives around testing and verification.
otterley 5 hours ago
autoexec 7 hours ago
bunderbunder 9 hours ago
But most white collar workers I know spend most their time either talking to each other, or thinking real hard about what to say the next time they get together to talk. If that’s generally true then it would suggest that there isn’t much white collar work that LLMs, which are now widely regarded as exceptionally poor communicators, can do particularly well.
pj_mukh 9 hours ago
The bull case (see: long term) is that the “organizing and talking” as long as it’s not something that is essentially human (like caretaking, nursing etc.) will happen between personal agents of sorts with only the deeper judgments being left to the humans opening up even more increases in productivity. This is already happening with PR reviews in a lot of workplaces.
Though I’m guessing most workplaces outside of tech will require total reorganization around this reality, but more likely this way of running will mean newer upstarts challenging incumbents refusing to change as it’s always been.
RussianCow 9 hours ago
LLMs can help tremendously with this part, though. Exploration/brainstorming, discovery, data analysis, etc can be made significantly easier with LLMs in most knowledge based fields.
alfjal 8 hours ago
I've seen much time wasted on strange ideas originated by an LLM. It's a lot easier for a human to organically have good ideas than to refute LLM-generated ideas that are a dead end.
epihelix 6 hours ago
All AI is doing here is rapidly identifying the people who have Peter-principled themselves.
> It's a lot easier for a human to organically have good ideas than to refute LLM-generated ideas that are a dead end.
What works better is when you can use an LLM to obtain a negative opinion on your own ideas that you think are good. LLMs aren't great at original thinking, but they're pretty good at finding the flaws in poorly thought-through human ideas.
bunderbunder 8 hours ago
The chatbots kept steering people toward uncreative ideas and overshadowing novel insights that might come from genuine domain expertise. AI generated meeting summaries became a force for miscommunication as they de-emphasized or even outright misrepresented what was really important about the conversation. And people became much worse at proving ideas out before making significant resource commitments thanks to effects that we now describe with terms like AI psychosis.
arthurcolle 8 hours ago
alfjal 8 hours ago
Some companies are releasing code that nobody on the owning team has ever read or reviewed, or has any inkling exists.
arthurcolle 6 hours ago
tdhz77 6 hours ago
vips7L 4 hours ago
shermantanktop 5 hours ago
sho 3 hours ago
I'm not sure you've quite thought this through. "Holding it wrong" is very, very often the case, especially with technology.
4ff4 5 hours ago
euiei 9 hours ago
Sigh. Where are the actual smart people here who think deeply? This surface level stuff doesn’t cut it
spunker540 4 hours ago
triceratops 3 hours ago
The god machine can prompt itself. Just go "Write me a prompt to <whatever>. Ask me as many clarifying questions as needed.". It is not a differentiator.
Deep domain knowledge and pattern matching that allows you to quickly verify outputs might be a differentiator. Maybe. But you might end up the bottleneck if you verify everything.
Aerroon 6 hours ago
giovannibonetti 8 hours ago
bunderbunder 7 hours ago
All of it relies on learning and information processing capabilities that LLMs simply do not have. The idea that any of it will automatically emerge from the current architectures simply by adding more parameters is about as sensible as thinking that if you keep counting long enough eventually you’ll get to complex numbers. We’re going to need to invent new technology to fill these gaps, and there is no way of knowing how long those developments will take.
bluefirebrand 2 hours ago
euiei 9 hours ago
When this stuff implodes it’s gonna be hilarious.
The bankers thought they were really smart in 2008. When people get carried away and take it too far things will fall apart.
mancerayder 5 hours ago
lokar 6 hours ago
coldtea 8 hours ago
Maybe that's just something which made sense when they were necessary to do the not yet automatable informattion processing work, but that only took like 20% of their time, so they added meetings, and useless work to keep them busy.
ashdksnndck 6 hours ago
I want to believe reasons why this won’t work, but it feels like cope. It would be smart to be prepared for the world where knowledge workers get replaced by AI and also prepared for the world where AI is a bubble that pops.
The goalposts on what neural nets can possibly do have already moved past chess, go, protein folding, coding, poker, math, the Turing test…
4ff4 5 hours ago
lol aka you have zero conviction about what you are saying.
So whats the point of you posting? Zero value add. In fact bordering on negative value.
ashdksnndck 4 hours ago
You want to drive across town. You, only willing to deal in certainties, must decide ahead of time whether you will get in a crash. If you determine won’t get in a crash, you make the drive, pedal to the metal. If you determine you would crash, you stay home.
Me, willing to deal in probabilities, make the drive at the speed I feel reasonable and wear a seatbelt.
stefan_ 9 hours ago