I wouldn't go back, regardless of salary offer, unless I didn't have any other jobs lined up. If I'm not employed than any job (even a bad one) beats being unemployed. But if I was employed, I wouldn't go back to a job where they laid me off for stupid reasons, no matter how much money they offered.
When you first lose everything, in the process you end up having to pawn expensive principles like that, so when other things like this happen, it's easy to seize the opportunity.
This is going to be the norm across the board as the models have failed to live up to the hype.
I do think LLMs and agents and all are great at helping you through tough problems but we aren’t there yet on getting them to do all the work while we just architect and design. Again, it’s close, and for your use cases you might be there already but for low level and big corporate lift and shifts, it’s not there yet.
I have agents, agents of agents, and I still find myself having to carve big chunks of my project off and feed it to the dogs because it’s garbage code. (GLM-5.2)
Documentation driven development is your friend here. 75% of my workflow is generating documentation, at ever lower levels of abstraction, until it’s just code. The code usually comes out optimal, clean, and bug free (after passing tests) and. Suuuuuper well documented lol.
> 75% of my workflow is generating documentation, at ever lower levels of abstraction, until it’s just code
Some might hate that writing code (which they enjoy) is turning into that, others might doubt the efficacy of doing that and the claims about it working so well.
Personally, I’d say that docs help as long as they’re meaningful and not too long (even AI tools have limited context), but you probably also want to codify what you can into code.
For example I wrote a tool in Go and goja called ProjectLint (not public yet but anyone can do that in a week) where you write custom rules in regular ECMAScript that can check whatever you want - code conventions across languages, project structure and architecture and all the stuff that goes under “In this project, we do X but don’t do Y” that just telling an LLM about (or colleagues) will be worth nothing (even memories and focus are limited), instead CI gates that.
I guess I reinvented a simplified and stack-agnostic version of ArchUnit but whatever, it works for me and I can use the same tool in Python and Java projects and elsewhere as well as parallelize all the read only checks and run sequentially the potential-write ones that might auto-fix stuff.
I’m sure it depends on the project, stack, and dev. I know loc is a terrible metric, but …
For me, my human only productivity in the firmware work I do is usually around 100-500 loc a day on good days. Obviously more when clean-slating the initial work on a project , but that’s typically a day or two and the same ratios apply.
With ai tools, I roughly 4x that with the same effort, or 2x it working lazily from my phone playing with my 2 year old.
The code is typically also more compact so the LOC metric is strong here IMHO.
Overall I have about the same number of bad-unproductive days, far less bugs (but worse bug hunts) and 10x better documentation lol.
ProjectLint sounds like an excellent tool for LLMs to use! (A tiny bit is sarcasm here) but seriously, delegation of (flagging) decisions off to deterministic tools is exactly the right call whenever it’s practical to do so. We write a lot of tools for just that, often single use python scripts.
I tend to feel like, I start out with a rough idea of a program I want to write in my head. I find it easier to just write the code directly than to write a document with sufficient detail about how I want it to work for an LLM to actually write the right code, then have the LLM write the code. And the resulting documentation is about as likely to be useless or a burden as it is to be helpful in the future.
The dream of perpetual labor machine is something capitalists are willing to destroy the planet in order to chase their fictitious dream. Oppressors must be stopped.
There was a reason why the phrase "perpetual" was used, to invoked how unrealistic perpetual energy machines are and how futile it is for the human race to chase such dreams.
How many tens of trillions of capital have been incinerated in reducing the quality of life for workers compared to actually uplifting them?
Sure, you'll need energy inputs, you're not going to beat thermodynamics. But we're not capturing even 0.01% of available energy yet, there's a lot of room to grow.
Industrial capitalism has been fantastic for quality of life. Here I am sitting in an air-conditioned office browsing HN during a workday, instead of slaving away in the fields as a peasant farmer. I'll take more automation please.
I was wondering the other day why we didn't put this level of effort into building a highway across the Atlantic and the Pacific. It seems to me if we just piled bricks made with all the money dumped into AI in the ocean, we could easily have done this. Likewise we could have just build a canal across the United States from the Atlantic to the Pacific. These efforts would have drastically reduced shipping costs and risks but they look impossible (and stupid) on paper so no one tried them.
Why is AI different?
Because it happens in a computer and many people think that makes something easy, like CGI or computer hacking in movies. It's intangible magic and belief is the product sold to investors.
> The return of the veteran engineers at Ford cuts against the prevailing wisdom — and fear — that AI will replace all kinds of knowledge workers. But Ford found the machines couldn’t replace experience.
I'm not sure this story is illustrative of that, when you have a VP of engineering saying “Over prior years, we didn’t pay as much attention as we should have to the experience of our most knowledgeable engineers that have been with us through many product cycles.”
He's saving face while almost certainly trying to figure out how to make the new systems work so that next time he won't need to rehire engineers.
> He's saving face while almost certainly trying to figure out how to make the new systems work so that next time he won't need to rehire engineers.
Yup. They jumped the gun. Now they need to hire them back so they can loot their expertise and never hire another senior. I'm not saying this will work, but it's pretty obviously the plan.
Maybe. That's one interpretation. A lot of hiring/firing decisions get read through the lens of AI, hard pro or hard con. Reality is always a mixed bag. They certainly will want to try to build up a better automated pipeline, but the question is can they, and can they cost-effectively vs hiring a few more people?
Pre-AI version: Oops, you laid off the higher-salaried people without having them train their replacements, so bring them back, long enough to do that.
Now, that training[*] will be for both AI models and lower-salaried hires.
Perhaps a second mistake by those who thought they didn't need their most experienced people: Now they think they just need to train the AI better, and then new-grad "AI native" hires will be the most cost-effective way to operate/oversee the AI and do whatever it can't.
[*] edit: originally typed "replacement" when I meant to type "training"
Is there any substantial number of companies actually training AI? Or do you count writing skills files for Claude as "training"? (Cause it really isn't..)
Well for grandma on the street I can accept that, but shouldn't at least the tech community be more precise in terminology? "AI" is also a misnomer. So many things in our industry are that it always takes some layers of digging in a new area to understand what they actually mean because the words have shifted their meaning.
I intended for the entire sentence to be in terms of the thinking of top leadership.
And to gloss over how that improvement would actually happen. (Not knowing what they've currently done and want to do, but for example, guessing: probably in partnership with vendors, consultants, etc., iterative and experimental process and tools improvements, and involving a variety of approaches and refinements.)
And for people focusing too much on AI, Xiaomi kicked their first vehicle into production with a fully automated factory three years ago [0]. That's where the industry is going and has tried to go for decades now.
They might want to also reduced head out on the designing side, but it's also an ongoing trend that started before the AI boom.
That's not an industry that will keep hiring as much as they did in the past, however it turns out.
> Mistakenly we thought that by just introducing artificial intelligence and ingesting the design requirements that we had, that that would produce a high-quality product.
Clearly a lot of careful thought went into their strategy of using AI and firing engineers.
Had a couple of Taurusii back in the day. 100% ended up having a problem where the power steering pump shit the bed because a plastic piece in the pressurized side failed. Paid to repair one, oem pump broke on drive home due to same plastic piece under pressure.
My point being, Ford's had shit for brains for decades. Its a fucking wonder any of their vehicles make it out of the parking lot.
I had a Focus in the 2000's that was the most reliable car I ever owned. Rust got it eventually but it still started instantly at any temperature and ran like a new car.
LOL yeah I had that too. Forgot about it. Cheap fix was an aftermarket door handle from Amazon or RockAuto or someplace like that.
I'm not saying it was a perfect car. The interior was cheap, the sheet metal seemed to be recycled tin cans, and it definitely showed its age by the time I got rid of it. But that engine and drivetrain seemed to be bulletproof.
It was also designed by European engineers, not in Michigan. Not saying that's the reason the Focus is more reliable than a Taurus but they didn't follow the "typical" Ford design process at the time for that vehicle. For what it is worth I owned a 1992 Taurus and it left me stranded more times than I can count. Just some of the issues I had were a water pump that exploded and a seized A/C compressor.
Pretty much everything Ford brings to the US that was designed in Europe is loathed by anyone who has to own it out of warranty.
Turns out that when you have a building full of engineers in Germany or England their domestic engineering culture results in work output not all that different from the sort of stuff people chastise BMW and Land Rover for.
That said, the Escort, and to a lesser extent the Focus, are generally considered very good vehicles.
This idea is everywhere right now, that AI is some magic black box that will solve all your business problems. The sentiment is spreading through the exec team where I work now too. It's like a disease.
C-suites completely disconnected from reality and assuming we've already achieved ASI/AGI, and marketing teams & business journals are only furthering that narrative.
It's so weird. I don't know what it is about AI that causes people to throw all thought and caution to the wind and charge forward blind. Its like they've been chomping at the bit for decades to get rid of those pesky humans and are so hyped up over it they can't see clearly anymore.
> It's so weird. I don't know what it is about AI that causes people to throw all thought and caution to the wind and charge forward blind. Its like they've been chomping at the bit for decades to get rid of those pesky humans and are so hyped up over it they can't see clearly anymore.
It's just a hype cycle. In my 15 years in data, I've seen around 3-4. Every time leadership get way too invested in the possibilities, and they waste tons of money on doomed efforts. A good example of the prior one was "Big Data" which was even more pointless than the current AI boom.
Don't get me wrong, there is valuable tech there (at the very least, being able to reliably generate structured data from unstructured input is incredibly valuable in data), but the current hype is way off the charts.
I think you are misleading people by calling it a "hype cycle". There is no going back from this technology. It is going to encroach every part of lives more and more.
What does hype even mean concretely? I think this is just a coping mechanism if you ask me.
Hype cycle doesn’t imply the technology has no value. But we should be able to talk about it as the boring, nerdy technology it is without that whole doom trolling and “AI will literally solve everything”
Er, what? Intricacies of a transformer pipeline might be boring and nerdy, but the results are not. BTW, I've yet to find any strong argument on why the current ML approaches are bounded below the level you find appropriate to be bored.
The idea is there’s a rush of irrational exuberance when an “innovation trigger” makes a new toy looks promising, and everybody rushes to use it for everything, regardless of whether its suitability-for-purpose is proven. Inevitably many of those pioneers find that it’s not good for their particular problems after all; usage reaches a “peak of inflated expectations,” and crashes into a “trough of disillusionment.”
Then the tech enters a quieter and more gradual “slope of enlightenment” as people work out use cases where the tech actually adds value; then adoption reaches a “plateau of productivity.”
Worth a glance at the way they map this to prior waves of technological exuberance.
From your video, it looks like your definition of hype involves a situation where eventual adoption increases above what is in the hype today.
Here's what the parent comment thinks:
> It's just a hype cycle. In my 15 years in data, I've seen around 3-4. Every time leadership get way too invested in the possibilities, and they waste tons of money on doomed efforts. A good example of the prior one was "Big Data" which was even more pointless than the current AI boom.
Obviously the parent doesn't think of hype the way you think of it because they claim that big data was pointless -- they don't see the eventual "slope of enlightenment". They think of hype cycle in the colloquial way and I was responding to that.
I see this all the time in the website and frankly the patronising "but actually hype means something else" is pointless and pedantic. I urge you to respond to words within the context and not bringing in academic definitions.
The person I replied to put words in your mouth. You and I agree what you meant. You mean that the hype would die down and won’t come back up again. Ever. So reply to the person above who thinks you mean hype this way.
AI is particularly infectious among C suites, because AI is great at spewing words. A substantial portion of folks in those positions are there because of family connections, existing wealth, etc., and their only contribution to the business is similarly spewing words. They went to good colleges where they excelled at spewing words. They worked cushy / hard jobs where they had to spew the just the right normal predictable words for this context, perhaps at a large volume and with little notice... and the words were hard words... not known to those outside the industry.
For those that lack initiative, strategy, a real understanding of their business, engineering, etc., the spewing words is the whole thing. It overshadows their entire understanding.
> Its like they've been chomping at the bit for decades to get rid of those pesky humans and are so hyped up over it they can't see clearly anymore.
This is precisely it. Here's my analysis:
AGI is a savior figure for the capitalist class. A tech version of the Second Coming, delivering them from the pesky demands of workers, like a living wage or (gasp!) sick leave.
That's why they're all so obsessed with it, it has religious-ideological component to them. When you hear them talk about AGI, there's always this weird eschatological vibe with it.
Unfortunately, they're blinded by their beliefs and can't think things through even one step further. Even if their cyberjesus comes down to them through the machine and replaces all workers, who's gonna buy all their stuff then?
All they're doing in their capitalist zealotry is ringing in the end of capitalism.
These guys have squeezed out every cost and slack from their system. They've found the exact revenue-maximizing prices and segmentation for their products. They've cut quality to the point where customers will just barely not reject their product. They have used every legal and accounting trick at their disposal to keep that line going up. But, next quarter, line must still go up!
The final massive cost to cut are all those damn human bodies that they they still have to keep around. They've driven down salaries and benefits to the minimum they can get away with, and they've extracted the maximum value from employees they can. But they haven't figured out how to get rid of them entirely. They are staring down the barrel of the gun and just can't see a way to cut this cost further. Now, magic AI comes along, and everyone is saying that the black box can replace those bodies. The C-suites believe it. They have to believe it. Line must go up! This is how they'll do it for a few more quarters. This is why the messaging is so unified across the industry, across every C-suite out there. They all need to believe.
> Line must go up! This is how they'll do it for a few more quarters. This is why the messaging is so unified across the industry, across every C-suite out there. They all need to believe.
The real danger for the economy is when the runway finally runs out. And I believe we are at a perfect-storm scenario... AI is obviously a giant wash-trading bubble that alone would be sufficient to trigger a repeat of the 2007ff crisis. But on top of that, we got the issue you mentioned, i.e. everyone running out of kool-aid and noticing it too late, with no easy way of turning around, and we got the war risk and supply chain shocks thanks to Iran and Russia, and and and.
And that's how you get a new war. Line must go up isn't only for the corps, its for US GOV debt too. Interest payments are already close to $1 trillion. As soon as GDP growth doesn't stay ahead of the compounding interest, the music stops. The line must go up or you get a sovereign deb crisis. When all other avenues are expended, the gov must force the economy to expand by any means necessary. Historically, that meant war.
Government debt can go down too. Democrat governments have been historically pretty good at balancing the budgets - only for their Republican successors to waste all of the effort on tax cuts for the rich and, yes, yet another dipshit war.
> I don't know what it is about AI that causes people to throw all thought and caution to the wind and charge forward blind.
1. Zero personal risk because cargo culting is a valid excuse in Executive World. If investors are on board, its good, no matter how stupid or destructive it actually is.
2. Top leadership's friendship with the country's leadership equals access to cheap debt financing since money is all fake and generated out of thin air
> It's so weird. I don't know what it is about AI that causes people to throw all thought and caution to the wind and charge forward blind.
My favorite theory about this is that we're all used to "speech == intelligence" and now that we have something that can produce coherent speech, it seems like it must be intelligent to people who don't know how it works. Even people who know how it works still anthropomorphize it to a weird degree. So a business person sees this thing that's both intelligent (to them) and superhumanly fast and it seems like the ultimate silver bullet.
Earlier this year I’ve been in calls with leaders from top US companies where their strategy was basically “we have to switch absolutely everything to agentic right now, otherwise we are dead”. That was the full thought.
That made reading their subsequent layoff blog posts pretty depressing
Nice thing about the C-suite is that you get authority and compensation without responsibility. You just claim responsibility when things are good. And when bad, the underlings who have responsibilities but no authority take the heat.
Personally, for me that represents job security. Having a human with a high level of domain knowledge in the loop seems pretty required to get any meaningful results.
solution is to always do what it has seen in training data and how it was RL. But ai companies dont tell you that. so you have to reverse engineer its training and stick to to that.
These are no general purpose machines. They are shipping a subset mindset not general intelligence like they want us to belive .
A friend of mine prepared an arsenal of hooks and the like to address this and LLMs still disobey them at times.
It's a model of language, yes? Trained on a big corpus of text.
I have read a lot of stories and accounts in which people were told not to do something and inevitably they did it. Like, lots. Far more than stories and accounts in which people were told not to do something and they then didn't do it.
If I'm reading a story or account of something, and it's really hammered home that they've been told not to do something, it's kind of inevitable that they will then do that. I'm not even an LLM and I noticed that's the way these things usually go.
So is an LLM just doing what it's been trained to do? Sometimes in the stories and accounts, there's a whole lot of time and tension before the bad thing happens, but that's just part of the fun.
We didn't have to do that. It is, in fact, extremely stupid that we have done that. Computers are valuable because they are fast and deterministic. Fast but stochastic has no value.
I don't understand how that angle keeps surviving. It is in the interest of the rich and powerful to keep the vast majority of society in jobs and pay them a wage. That's what they use to consume the things that drive the economy which ultimately makes the rich richer. The narrative that the rich want to get rid of workers is as nonsensical today as it was decades ago when I heard it the first time. It doesn't make any sense.
Those employee wages for a product is a 20th century way of making money. Taking investor cash and paying it back with supplier "investments" is how "capitalism" works in today's economy. The labor market and products is just the money laundering cover story for ponzi schemes. It's way faster and more lucrative taking money from the rich in big chunks than taking it from the poor in teensy amounts. This is why everything sucks now, no one cares about the product.
How interesting. So a Ford car is now more reliable than a Toyota soon after purchase but Toyota didn’t fire anyone and Ford fired, implemented automated reviews, and rehired. So their process didn’t bring them back to neutral. It placed them above the traditionally reliable manufacturers.
So maybe the key is firing everyone and then rehiring the good guys after you implement automated systems.
Though I’m somewhat surprised. I didn’t expect Porsches to top a reliability measure. I thought they were in the “fancy but unreliable” bin. Interesting.
This seems like a totally crazy statement. The only common thing that a current and 50-year-old 911 share is that there are six holes in the engine block.
I've had two different Porsches, a Cayman S and a Macan. Neither gave me a day of trouble. You just have to do all the maintenance, which is obviously expensive.
porsche is part of volkswagen, so it's not that surprising that they're decently reliable. i probably see 10 porsches for every ferrari, lamborghini, etc that i see, and i think a large part of that is reliability - even absurdly rich people don't want to deal with an unreliable car when there is a more reliable alternative.
I wonder if Porsche is allowed to exist in point where they are not fully cost optimised so there is more spend on those slight things that keep reliability. Most other large manufacturer cars seem to be cost optimised while least amount of that is carried over to customers...
Setting aside how shortsighted it is to fire your employees to replace them with AI, Ford also screwed up by firing the wrong employees. LLMs work best in the hands of experienced senior engineers who can work at a high level of abstraction because they already understand all the pieces underneath.
In a sense, using an LLM agent is like providing instructions to a very smart, very quick junior who despite being brilliant has some blind spots and lacks institutional knowledge. That's something that seniors excel at, so by firing your seniors you've fired the people best positioned to make full use of LLMs.
"Ford rehires 350 engineers after AI fails to preserve expertise or train juniors" In order to rehire someone they must laid off or fired? You don't rehire new employees?
And the verge is covering it too:
https://www.theverge.com/transportation/956316/ford-quality-...
It's just so strange any other profession have unions or bodies that protect their job against this sort of practice.
if software devs were lawyers then AI would've been banned
If the company tries to layoff 10% "due to AI" the remaining 90% can strike.
History is full of union solidarity vs idiotic management.
I do think LLMs and agents and all are great at helping you through tough problems but we aren’t there yet on getting them to do all the work while we just architect and design. Again, it’s close, and for your use cases you might be there already but for low level and big corporate lift and shifts, it’s not there yet.
I have agents, agents of agents, and I still find myself having to carve big chunks of my project off and feed it to the dogs because it’s garbage code. (GLM-5.2)
It’s human in the loop over and over again tho
Some might hate that writing code (which they enjoy) is turning into that, others might doubt the efficacy of doing that and the claims about it working so well.
Personally, I’d say that docs help as long as they’re meaningful and not too long (even AI tools have limited context), but you probably also want to codify what you can into code.
For example I wrote a tool in Go and goja called ProjectLint (not public yet but anyone can do that in a week) where you write custom rules in regular ECMAScript that can check whatever you want - code conventions across languages, project structure and architecture and all the stuff that goes under “In this project, we do X but don’t do Y” that just telling an LLM about (or colleagues) will be worth nothing (even memories and focus are limited), instead CI gates that.
I guess I reinvented a simplified and stack-agnostic version of ArchUnit but whatever, it works for me and I can use the same tool in Python and Java projects and elsewhere as well as parallelize all the read only checks and run sequentially the potential-write ones that might auto-fix stuff.
For me, my human only productivity in the firmware work I do is usually around 100-500 loc a day on good days. Obviously more when clean-slating the initial work on a project , but that’s typically a day or two and the same ratios apply.
With ai tools, I roughly 4x that with the same effort, or 2x it working lazily from my phone playing with my 2 year old.
The code is typically also more compact so the LOC metric is strong here IMHO.
Overall I have about the same number of bad-unproductive days, far less bugs (but worse bug hunts) and 10x better documentation lol.
Coding is definitely a different job though.
You just want to make sure you have it, and not your boss using it against you.
How many tens of trillions of capital have been incinerated in reducing the quality of life for workers compared to actually uplifting them?
Industrial capitalism has been fantastic for quality of life. Here I am sitting in an air-conditioned office browsing HN during a workday, instead of slaving away in the fields as a peasant farmer. I'll take more automation please.
Why is AI different?
Because it happens in a computer and many people think that makes something easy, like CGI or computer hacking in movies. It's intangible magic and belief is the product sold to investors.
I'm not sure this story is illustrative of that, when you have a VP of engineering saying “Over prior years, we didn’t pay as much attention as we should have to the experience of our most knowledgeable engineers that have been with us through many product cycles.”
He's saving face while almost certainly trying to figure out how to make the new systems work so that next time he won't need to rehire engineers.
Yup. They jumped the gun. Now they need to hire them back so they can loot their expertise and never hire another senior. I'm not saying this will work, but it's pretty obviously the plan.
Now, that training[*] will be for both AI models and lower-salaried hires.
Perhaps a second mistake by those who thought they didn't need their most experienced people: Now they think they just need to train the AI better, and then new-grad "AI native" hires will be the most cost-effective way to operate/oversee the AI and do whatever it can't.
[*] edit: originally typed "replacement" when I meant to type "training"
And to gloss over how that improvement would actually happen. (Not knowing what they've currently done and want to do, but for example, guessing: probably in partnership with vendors, consultants, etc., iterative and experimental process and tools improvements, and involving a variety of approaches and refinements.)
And for people focusing too much on AI, Xiaomi kicked their first vehicle into production with a fully automated factory three years ago [0]. That's where the industry is going and has tried to go for decades now.
They might want to also reduced head out on the designing side, but it's also an ongoing trend that started before the AI boom.
That's not an industry that will keep hiring as much as they did in the past, however it turns out.
[0] https://youtu.be/v6jb6PP4APc
Clearly a lot of careful thought went into their strategy of using AI and firing engineers.
My point being, Ford's had shit for brains for decades. Its a fucking wonder any of their vehicles make it out of the parking lot.
I'm not saying it was a perfect car. The interior was cheap, the sheet metal seemed to be recycled tin cans, and it definitely showed its age by the time I got rid of it. But that engine and drivetrain seemed to be bulletproof.
Pretty much everything Ford brings to the US that was designed in Europe is loathed by anyone who has to own it out of warranty.
Turns out that when you have a building full of engineers in Germany or England their domestic engineering culture results in work output not all that different from the sort of stuff people chastise BMW and Land Rover for.
That said, the Escort, and to a lesser extent the Focus, are generally considered very good vehicles.
k
In that order, apparently.
Step 1: 30 minute conversation with AI on how to use AI. Step 2: fire everyone.
Step 3: Rehire key personnel at lower cost than whomever was fired in Step 1. Step 4: Take credit for cost reductions . . . and give yourself a raise!
C-suites completely disconnected from reality and assuming we've already achieved ASI/AGI, and marketing teams & business journals are only furthering that narrative.
It's so weird. I don't know what it is about AI that causes people to throw all thought and caution to the wind and charge forward blind. Its like they've been chomping at the bit for decades to get rid of those pesky humans and are so hyped up over it they can't see clearly anymore.
It's just a hype cycle. In my 15 years in data, I've seen around 3-4. Every time leadership get way too invested in the possibilities, and they waste tons of money on doomed efforts. A good example of the prior one was "Big Data" which was even more pointless than the current AI boom.
Don't get me wrong, there is valuable tech there (at the very least, being able to reliably generate structured data from unstructured input is incredibly valuable in data), but the current hype is way off the charts.
What does hype even mean concretely? I think this is just a coping mechanism if you ask me.
Er, what? Intricacies of a transformer pipeline might be boring and nerdy, but the results are not. BTW, I've yet to find any strong argument on why the current ML approaches are bounded below the level you find appropriate to be bored.
https://www.gartner.com/en/research/methodologies/gartner-hy...
The idea is there’s a rush of irrational exuberance when an “innovation trigger” makes a new toy looks promising, and everybody rushes to use it for everything, regardless of whether its suitability-for-purpose is proven. Inevitably many of those pioneers find that it’s not good for their particular problems after all; usage reaches a “peak of inflated expectations,” and crashes into a “trough of disillusionment.”
Then the tech enters a quieter and more gradual “slope of enlightenment” as people work out use cases where the tech actually adds value; then adoption reaches a “plateau of productivity.”
Worth a glance at the way they map this to prior waves of technological exuberance.
From your video, it looks like your definition of hype involves a situation where eventual adoption increases above what is in the hype today.
Here's what the parent comment thinks:
> It's just a hype cycle. In my 15 years in data, I've seen around 3-4. Every time leadership get way too invested in the possibilities, and they waste tons of money on doomed efforts. A good example of the prior one was "Big Data" which was even more pointless than the current AI boom.
Obviously the parent doesn't think of hype the way you think of it because they claim that big data was pointless -- they don't see the eventual "slope of enlightenment". They think of hype cycle in the colloquial way and I was responding to that.
I see this all the time in the website and frankly the patronising "but actually hype means something else" is pointless and pedantic. I urge you to respond to words within the context and not bringing in academic definitions.
I think the tech is useful, but the hype is ridiculous so I expect lots of companies to have large share price declines et al when this settles.
Big data also had a good core, but all the e-commerce sites building a data lake wasted lots of money.
For those that lack initiative, strategy, a real understanding of their business, engineering, etc., the spewing words is the whole thing. It overshadows their entire understanding.
> Its like they've been chomping at the bit for decades to get rid of those pesky humans and are so hyped up over it they can't see clearly anymore.
This is precisely it. Here's my analysis:
AGI is a savior figure for the capitalist class. A tech version of the Second Coming, delivering them from the pesky demands of workers, like a living wage or (gasp!) sick leave.
That's why they're all so obsessed with it, it has religious-ideological component to them. When you hear them talk about AGI, there's always this weird eschatological vibe with it.
Unfortunately, they're blinded by their beliefs and can't think things through even one step further. Even if their cyberjesus comes down to them through the machine and replaces all workers, who's gonna buy all their stuff then?
All they're doing in their capitalist zealotry is ringing in the end of capitalism.
Knowledge or skilled workers can be used by the AI for swarm training data generation; what value do the execs have to AI?
I think the most beautiful part of capitalism is selling elites rope to hang themselves.
These guys have squeezed out every cost and slack from their system. They've found the exact revenue-maximizing prices and segmentation for their products. They've cut quality to the point where customers will just barely not reject their product. They have used every legal and accounting trick at their disposal to keep that line going up. But, next quarter, line must still go up!
The final massive cost to cut are all those damn human bodies that they they still have to keep around. They've driven down salaries and benefits to the minimum they can get away with, and they've extracted the maximum value from employees they can. But they haven't figured out how to get rid of them entirely. They are staring down the barrel of the gun and just can't see a way to cut this cost further. Now, magic AI comes along, and everyone is saying that the black box can replace those bodies. The C-suites believe it. They have to believe it. Line must go up! This is how they'll do it for a few more quarters. This is why the messaging is so unified across the industry, across every C-suite out there. They all need to believe.
The real danger for the economy is when the runway finally runs out. And I believe we are at a perfect-storm scenario... AI is obviously a giant wash-trading bubble that alone would be sufficient to trigger a repeat of the 2007ff crisis. But on top of that, we got the issue you mentioned, i.e. everyone running out of kool-aid and noticing it too late, with no easy way of turning around, and we got the war risk and supply chain shocks thanks to Iran and Russia, and and and.
1. Zero personal risk because cargo culting is a valid excuse in Executive World. If investors are on board, its good, no matter how stupid or destructive it actually is.
2. Top leadership's friendship with the country's leadership equals access to cheap debt financing since money is all fake and generated out of thin air
3. Too big to fail
My favorite theory about this is that we're all used to "speech == intelligence" and now that we have something that can produce coherent speech, it seems like it must be intelligent to people who don't know how it works. Even people who know how it works still anthropomorphize it to a weird degree. So a business person sees this thing that's both intelligent (to them) and superhumanly fast and it seems like the ultimate silver bullet.
That made reading their subsequent layoff blog posts pretty depressing
I don't have high hopes that there exists a bulletproof solution to this.
These are no general purpose machines. They are shipping a subset mindset not general intelligence like they want us to belive .
It's a model of language, yes? Trained on a big corpus of text.
I have read a lot of stories and accounts in which people were told not to do something and inevitably they did it. Like, lots. Far more than stories and accounts in which people were told not to do something and they then didn't do it.
If I'm reading a story or account of something, and it's really hammered home that they've been told not to do something, it's kind of inevitable that they will then do that. I'm not even an LLM and I noticed that's the way these things usually go.
So is an LLM just doing what it's been trained to do? Sometimes in the stories and accounts, there's a whole lot of time and tension before the bad thing happens, but that's just part of the fun.
He valuations of a bunch of AI unicorns disagree.
So maybe the key is firing everyone and then rehiring the good guys after you implement automated systems.
Though I’m somewhat surprised. I didn’t expect Porsches to top a reliability measure. I thought they were in the “fancy but unreliable” bin. Interesting.
An expensive process.
In a sense, using an LLM agent is like providing instructions to a very smart, very quick junior who despite being brilliant has some blind spots and lacks institutional knowledge. That's something that seniors excel at, so by firing your seniors you've fired the people best positioned to make full use of LLMs.