The answer has nothing to do with rationing. When all software is designed to grab the most attention, serve the most adds, collect the most data, deliver the maximum amount of the cheapest content and generally be as exploitive of its users as possible, no amount of rationing or not rationing will effect it.
McMaster Carr, a company worth a little over a billion, sells metal, bolts, casters, and many other industrial bits and bobs. Their website is infinitely better than Amazons. Why? It has nothing to do with rationing and everything to do with trying to actual deliver a good customer experience.
marcus_holmes 20 hours ago [-]
I'm even more cynical; most software from large organisations is bad because it was never designed for the user, it was designed to make the product manager's career advance.
Two factors:
- middle management do not have contact with either customers or shareholders. They report to managers, and only have managers reporting to them. They're insulated from the actual consequences of their decisions.
- the only way to get recognised for promotion is to create change. Any change. Doesn't matter if it's bad (blame can be shifted), it just needs to be noticed.
If a feature of the product hasn't changed for a long time (because it works for the user), it becomes more of a target for changing. If it changed recently, it creates less impact if it changes again, so isn't a good target. So the product becomes worse over time, not better.
sanketsaurav 46 minutes ago [-]
Agree that incentives explain a lot of why software is bad. But I don’t think they explain why it all looks the same. For instance, every code forge copied the pull request, including the ones with no ads to serve.
xboxnolifes 21 hours ago [-]
Ironically, so many modern tech companies are arguably less incentivized to provide a good technology experience than non-tech companies, since the tech company's value/revenue primarily comes from directly exploiting their users for their data and/or attention. Improving the user experience frequently directly reduces their revenue. No amount of free LLM usage can allow a video serving company to both serve ad-free videos and make money from ads in their videos.
Though, maybe the irony only exists because we use the broad term tech company instead of the more generally accurate terms data harvesting company and attention consuming company.
NegativeK 20 hours ago [-]
Mcmaster also isn't (somewhat famously) competing on price. They won't rip you off, but you can't sort by price. There are no sales. With limited exception, they won't list brands on the site. (But if you contact them, they'll gladly tell you who so you can buy direct.)
Their target market is people who need to fix something quickly or get some industrial supply item tomorrow so work isn't stopped.
Massive corporations save money via the end user by taking a chainsaw to customer service. Large tech companies (and many other industries) will just quietly fuck the regular customer over because they can't profitably do customer service at scale. Even Amazon's effectively no questions asked return policy is a band-aid over the forgeries and trash that fill their hyperscale market.
If you have a problem with an item that Mcmaster sends you, they're going to go have a stern talk with the manufacturer. They sell reliability, not race to the bottom capitalism.
(I've also heard that they really take care of their employees. People in jobs that are automated away are always retrained; they'll pay for college with no qualifications, because education is good even if they don't directly benefit, etc.)
testing22321 8 hours ago [-]
> McMaster Carr, a company worth a little over a billion, sells metal, bolts, casters, and many other industrial bits and bobs. Their website is infinitely better than Amazons
I’ve bought a few dozen things from McMaster. For very competent people who know precisely what they want it is good, but I do find I spend an hour looking at the specs of “doodad” which will almost meet my needs when I stumble onto “doodaddy” which is a better fit. After I buy it, a friend points out they sell “doodada” which is even better.
Note I’m an engineer, but didn’t study in North America, so the names of things are different.
McMaster is terrible for product discovery, you have to know the precise name of the thing you want.
For the average person on the street Amazon is better.
kibwen 21 hours ago [-]
Crucially, McMaster-Carr is a private company, so their customers are the people giving them money for goods and services. As a public company, Amazon's customers are the gamblers on the stock market.
bemtevi 23 hours ago [-]
I don't think any of this makes much sense. The problem that this article ignores is that people build different things with AI, not the same things but faster. They build more superficial things, more throwaway things, things that are tailor-made for chasing the buck, and it just doesn't feel genuine or interesting.
It was much the same with search. Google search allowed us to find things faster at first...but over time websites responded to that and started doing SEO crap. Now almost every search results in a billion websites that say the same thing, but shallowly. Yes you can still find some gems but was it worth it?
AI makes things faster, but now it means we can target superificial needs very quickly...and so a new breed of crap software will be built with it. As long as there's a dollar sign at the end of the road.
tptacek 21 hours ago [-]
I don't understand this argument. Why do you think everyone is "chasing the buck"? That describes 0% of the work I've done over the last 6 months. There's no "bucks" in the music players, macro trackers, and source code review tools I've built.
watwut 12 hours ago [-]
Google search was enshittified intentionally. Its bad state is not the result of SEO, but consequence of a fact that more search queries means more ads money.
doginasuit 23 hours ago [-]
AI makes things faster, and it pushes the envelope on what can be built. Like a telescope lets the human eye see further into space, an LLM helps track more ideas and structures within working memory. Yes, people will use this new potential to build slop. And other people will use it to build what has never been built before. The only question is which one we will prefer.
aselimov3 23 hours ago [-]
By definition an LLM is unable to build what has never been built before because you are going outside the realm of pretraining/RL. You might be able to combine components that already exist in a unique way, but even then you will be fighting against the mediocre making effect of LLMs
outworlder 23 hours ago [-]
That understates what LLMs are capable of doing. They can and do build things outside of their training dataset. As long as you can get there by following existing concepts, they can do it.
fwlr 22 hours ago [-]
Perhaps outside their training set in a strict sense of “this thing itself was not specifically in the training set” (although only perhaps - users often iteratively coax frontier models into achieving things just beyond their envelope, and those achievements could well be folded into the next model’s training set without you or I being aware of that). But even then, it is not directionally outside the training set: it steps outside not in the direction of the unknown, but in the direction of another part of its training set. An evocative metaphor would be dashing from cover to cover in the rain, because it lacks the umbrella or raincoat or simply the fortitude to tolerate getting wet that it would need to reach an uncovered destination.
doginasuit 59 minutes ago [-]
I see I didn't make this aspect very clear in my comment, but I believe that what I described is only possible when a human is in tight control. I agree that an LLM on its own will generate pure slop 9 times out of 10. As an extension of your own creativity, I believe it will push the boundary.
majormajor 18 hours ago [-]
That seems nonsensical. Consider a selection of even just 8 types of components or patterns that might be in a training set (or available to be found over Google and fed into a prompt by a harness, for that matter). There are 256 distinct ways to use 1-8 of them together.
And of course the number of libraries, known algorithms, etc, is much larger than 8. There's any number of ways to make something new by combining well-understood parts.
I don't think non-expert prompting is pushing the envelope on what can be built because I think the current generation of harnesses and models pushes to using well-understood patterns in well-understood ways but it's certainly moving the line for what has been built. And with more skilled prompting you can build pretty crazy stuff + extensive validation for it.
AlexCoventry 18 hours ago [-]
"By definition is unable to prove what has never been proved before because you are going outside the realm of pretraining/RL."
That’s a hard position to defend in the face of navier stokes
Tanjreeve 17 hours ago [-]
The problem where the centrepiece of the controversy was OpenAI admitting the potential of using private work happening by mathematician users as inputs?
kasey_junk 11 hours ago [-]
Your argument is that advanced mathematics, that did not solve navier stones, were in the weights and that enabled them to solve it. And you think it butresses the argument that the llms can’t do _coding_ tasks they haven’t seen before?
Tanjreeve 9 hours ago [-]
> Your argument is that advanced mathematics, that did not solve navier stones
Do you think playing dumb is making you look credible or something? The argument isn't "was maths in the dataset" the controversy was if someone with potentially working approaches narrowed the search space enough for the LLM not to fail
> And you think it butresses the argument that the llms can’t do _coding_ tasks they haven’t seen before?
This isn't an argument being made. At least try to respond to what's actually written rather than making yourself look foolish and trying to wordplay your way to a point.
kasey_junk 8 hours ago [-]
> By definition an LLM is unable to build what has never been built before because you are going outside the realm of pretraining/RL
That’s pretty hard to read differently. Can you add the nuance I’m missing?
Tanjreeve 7 hours ago [-]
Well being a different person seems quite key. Not that I 100% disagree with the full quote rather than your selective quoting.
> By definition an LLM is unable to build what has never been built before because you are going outside the realm of pretraining/RL. You might be able to combine components that already exist in a unique way
kasey_junk 5 hours ago [-]
In most threaded Internet forums when you respond to a thread you are implying a continuation of the previous conversation.
It seems you weren’t following that norm and we were talking past each other. That’s fine and I’m happy to leave it at that.
Perhaps in the future you can recognize the local norms before resorting to personal attacks even if those norms are different than the parts of the internet you normally hang out in.
stillpointlab 23 hours ago [-]
I am not sure this is default a good thing.
I recall something I once heard about product marketing. The experiment was a product like jam, and they had 3 flavour options available in one case and 12 options available in another. It was a test on how hard it was to make a decision as well as how confident the person was they made the right decision later on.
One of Sartre's points on existentialism was that radical freedom leads to anxiety.
In some sense, AI is removing scarcity. It isn't clear to me that this is the same kind of thing as going from starvation to surplus. It is more like the jam case where we aren't increasing abundance in a total sense, we increasing the options within the existing surplus.
Those who haven't read Mythical Man Month are doomed to repeat it. I don't recall anywhere in the book where it states that the problems described in its essays don't apply to you if you're working with dozens of the most intelligent engineers in the world who can produce code in hours instead of days.
I was discussing this with someone else and used the analogy of nuclear energy. Although nuclear plants are highly complex and use forces that have only been understood for the last hundred years, at the end of the day it's still boiling water to create steam to spin a turbine, and nothing about nuclear innovation can engineer away the laws of thermodynamics.
Agents have and will continue to drastically change how we develop code, but it's my belief that drastically changing our software practices to speed up development just means we're going to hit engineering limits faster, rather than being able to overcome those limits entirely.
elcritch 21 hours ago [-]
I was telling a friend something similar last week.
Unless, or perhaps until, we figure out how to use LLMs effectively at the organization level we're still going to be limited to the "slime mold" level of organizational intelligence.
Will that happen? Most likely, but will it take a generation or ten? I suspect it'll be far harder for society to adapt at that level. We've had Mythical Man Month for decades yet software managers continue repeating the same mistakes.
Because of that the idea of OAI and Anthropic just moving to in-house product development makes sense.
Gigachad 21 hours ago [-]
At least what I’m seeing at work is that AI tools have generated thousands of security reports after scanning all of the old code. So our whole team has been bogged down for months now removing eval(), updating ancient packages, updating docker images that haven’t been rebuilt in years, etc. It shows no improvement to the end user but since all of these vulnerabilities have suddenly become very visible we have to action them all right now.
DenisM 20 hours ago [-]
I imagine this new awareness could create a lot of new jobs, given how much software out there should be in the same state.
I’m not sure if the job title tho. Software restoration?
Gigachad 18 hours ago [-]
I think this is just what many existing software devs have been busy with lately. The latest models are great at programming, but they have uncovered a massive amount of work that's somewhat time sensitive since you have to patch all of this before someone malicious runs the same scans and exploits it.
jimbokun 19 hours ago [-]
> Companies like Rippling and PostHog were the trailblazers, bundling several related products long before agents.
The trailblazer was Microsoft Office.
sanketsaurav 51 minutes ago [-]
Ha, good point. But Microsoft bundled it once it was already huge. Rippling did it as a startup, when that was considered a mistake. My bet is that agents make the Rippling path open to far more companies.
xivzgrev 24 hours ago [-]
People respond to incentives: Why do a hard thing when I can do an easy thing?
During the hype cycle, making "x for y" is easiest to get funded and exit.
Once the hype crashes and we're in the trough of disillusionment, then people are bearish. So founders either move to the next hype cycle or figure out a more durable business from the ashes
cpgxiii 19 hours ago [-]
I don't think the last decade of tech (particularly SV) hiring can in any way be defined as "rationing". If anything, we have seen large software companies grossly over-hire for all of the expected dumb reasons (mostly internal politics) and find themselves in the very predictable situation that more people != more productivity.
sanketsaurav 47 minutes ago [-]
True for big tech in the last decade. But I think the rationing that mattered happened earlier, when these products were small. By the time they could afford to overhire, the shape of the product was already set.
jimbokun 19 hours ago [-]
My theory is that most people have more software in their lives already than they want or need.
TwelveEyes 24 hours ago [-]
To be honest I don’t believe this is the case. Admittedly, I don’t have any experience with the best of the best models, only up to clause Opus. AI accelerates coding for myself a hell of a lot, but frankly the amount of time I spend coding complex programs feels about the same. I think there’s still a bottleneck, and it has been reiterated here before: judgement, testing, and understanding. If AI really accelerated coding, no amount of cultural inertia would be able to hold it back. It’s just not.
Grombobulous 23 hours ago [-]
I was about to say something similar. The article’s premise is solid as long as we accept the assumption that AI is actually fundamentally changing the velocity and methodology of coding.
To me, so far it seems like it is and it isn’t simultaneously. Some things are going at insane paces. I see things like reverse engineering and emulation breaking barriers. But we don’t see things like the quality and complexity of highly architected systems getting much better, maybe even getting worse as management pushes velocity onto employees.
I think what this shows is that AI is great at tedium and brute forcing, but it’s not actually good at making substantially better software with less of that “rationed” expertise. At least, not yet.
ares623 23 hours ago [-]
All this talk about removing bottlenecks at all costs, and we haven't even considered the final boss of bottlenecks: the end users themselves. All the gains will be a waste if the end users have a choice. And as "The Goal" has taught us, we must remove the further-most bottlenecks first. Hence, we must remove all choice and power from end-users first.
Tell me I'm wrong.
24 hours ago [-]
ihumanable 6 hours ago [-]
> After Claude Code (AC), the economics that gave us every single piece of software we use today have inverted. Building software has become cheap, with the marginal cost of building a new feature approaching zero, while running it, now that AI sits in the critical path, has become expensive.
This is such a weird take, if I need some piece of software why would an ongoing operating cost be having an AI sitting in the critical path.
If the argument is that making software has become cheap because you can use AI to generate software, it doesn't seem to follow that you would build all software as some sort of shell around AI. If I ask claude to build me a task tracking app, I imagine it would build me something with some bog standard data storage and not decide that it needs to keep the tasks in an LLM.
I don't agree with most of this post, but this part stood out as supremely odd. There is a lot of software that isn't made any better by having AI jammed into it as an always on piece of the operating puzzle.
sanketsaurav 53 minutes ago [-]
> This is such a weird take, if I need some piece of software why would an ongoing operating cost be having an AI sitting in the critical path.
The inversion only applies to products where an AI model sits in the request path, which is a growing share of what’s being built. It definitely doesn't mean all software that will get built (and that wasn't my intent anyway), so maybe my initial sentence in the essay overreaches. I'll make an edit to address this.
marcus_holmes 20 hours ago [-]
I'm more interested in what the world looks like when everyone has their own personalised software.
taybin 19 hours ago [-]
Decoherance. What happens to open source if everyone’s stock is different?
marcus_holmes 19 hours ago [-]
Open Source itself relies on copyright, which doesn't apply to AI-generated code.
skinfaxi 11 hours ago [-]
Only code generated completely without human input I thought?
taybin 4 hours ago [-]
I mean, why would people collaborate or contribute?
jimbokun 19 hours ago [-]
Vast majority of people have absolutely no interest in this.
marcus_holmes 19 hours ago [-]
At the moment, yes, because "developing software" is still a thing that most people aren't interested in.
But if the answer to "hey Claude, I need to track my gym sessions, what's the best way of doing that?" is "I've built an app for your phone to do that", does that change?
01284a7e 23 hours ago [-]
"Build me a high speed railway. Make no mistakes."
Now that SaaS CRUD apps can be built in 10 minutes, we can maybe do something more worthwhile, no? So much wasted effort has gone into this field already, in my opinion.
PaulHoule 24 hours ago [-]
... by improving model efficiency so AI tokens are "too cheap to meter" or alternately not that much more expensive to operate than conventional sorts of software?
It's a different view of scaling. I mean a lot of people have this absurd idea of "super intelligence" which is like the cover of this book
i mean what does "intelligence quotient 10,000" even mean? On the other hand it's very easy to imagine a world where we have systems like the best ones we have now but we can really afford to use them.
cyberax 23 hours ago [-]
> i mean what does "intelligence quotient 10,000" even mean?
IQ tests measure, essentially, how fast you can do shallow pattern recognition.
And AI can do these kinds of tests easily.
Dylan16807 22 hours ago [-]
There are time limits but to say they're primarily measuring speed sounds wrong.
cyberax 15 hours ago [-]
Does it? Take it to the limit: would a single question from the IQ test with unlimited time be a serious challenge? Of course not. So time is the crucial part of the test.
My IQ is around 120, so it's higher than average, but nothing exceptional. But I'm pretty good at math, so I competed at the national level in my country when I was at school. And as a part of that, I got to attend a summer camp with people who have a REALLY high IQ. Several of them went to win the IMO.
They were able to solve problems much faster than I could during training sessions. Often by the time I started feeling that I had a way to the solution, somebody already had the complete answer ready to be discussed. But with unlimited time to think, I could solve most problems.
beloch 20 hours ago [-]
I think we're about to see a shift in terms of what is a viable software product and what isn't.
In the past, if you wanted to do a task, you had to choose from one of a few (or maybe just one) incumbents and make their tool fit your task. Failing that, build your own tool, but that was expensive. If you did build your own, the tool might have value and could potentially be sold to others.
Going forward, building your own will be much easier and cheaper. Even if the incumbent tools suit the task well, you might choose to build your own for a variety of other reasons. e.g. If the incumbent raises prices or chooses to enshitify. What you will probably not do is adopt some third party's vibe-coded clone of an incumbent's tool, because that won't be customized to your needs and it won't have the track record of the incumbent. You can't trust it or control it so it will be better to build your own.
If an incumbent exists and you can now vibe-code a clone, you probably can't sell that clone "as is" and expect to be successful. You might be able to sell your services to customize and refine tools for small businesses that can't really afford to hire a solitary soul to build their own solution, but don't expect that niche to persist long-term.
wonnage 18 hours ago [-]
Maybe OP can explain why with unlimited tokens neither OAI nor Anthropic have any first party apps other than the main Claude/chatgpt apps
There are only so many rust migrations you can do, at some point you gotta ship something user visible which means you have to deal with people which means the code was never the bottleneck even in 2014
tonymet 23 hours ago [-]
Before we were spending 10:1 on features:quality . I worry that ratio will remain, but at 100x or 1000x the velocity.
There's no reason to think that companies are suddenly going to be spending 1:1 features : quality because both are faster / cheaper .
So expect the low quality to continue, and naturally get worse, because complexity is going up 1000x
latentmirror 21 hours ago [-]
hehe...Before Claude(BC) and After Claude(AC).
In Claude we trust.
ares623 24 hours ago [-]
The free holiday will last forever. Make your plans accordingly.
9p 22 hours ago [-]
this reads like an ad.
mahirsaid 17 hours ago [-]
It seems the thinking behind why software exists in the first place revolves around two main questions: 1. What problem or bottleneck can society use help with, along with a large initial impact that can be further enlarged later. Are there attractions already out in the market: such as an echo amongst the general public adjacent to a clearer sentiment that can be heard. Mirror the reality that most of the greatest and grandest injection to market creators of a new market ( Mark Zuckerburg, Bill Gates, Jeff Bezos, Larry Page ) there many more you can name here in the U.S. and overseas, however the point is these are usually odd individuals that are not exactly a fit for society in general. Their views are often " outside looking in" to what appears to be a missing gap or a general vacuum. Mark zuckerberg was in college and the main goal was to to be inspired or posses an idea taken from someone else perhaps. Same goes for Jeff Bezos, who i might add had a comfortable job before taking the risk of selling books online, this could be a simple observation of his mother growing frustrated during the .com era where there might have been a stigma against physical books as the U.S. was going through a transition. There isn't really a clear path for most of these people just an idea that was given or inspired and an ability to "look in".
2. What is the end goal? The arena looks like that are venture capitalists, Co-founders, founders, angel investors. An MVP that can show off the minimal representation of your vision. At a time where these were steps that each serve their purpose and each come with their own challenges and journeys that are entirely different from one another.
There are oddities that currently exist without little echoed direction, there are two main operating systems for phones and computers. there are only a handful of tech companies with influence and control over direction to which the tech world heads towards. This was more true a few years back, however now we have tech companies directly involved with government and atrocities abroad. The tech companies you remember are no more along with it every piece of software. You might question the exaggeration and dramatized scene before you, but i assure you it is true. OpenAI was a disruptor until 13b was the aid that helped make them pop only after Elon Musk had something to do with the lab before their founded success. It is not that special anymore, just some tech company that comes to disrupt the market for a few years until that main hosting and dependence on the very firms you are looking to challenge. Web3 is not: A blockchain for the internet. A sickening redundant use of AI. Unintentionally reading articles on online of crime and most had section describing the criminal's activity and leading up to the crime with AI prompt they used to convict them. It is clear. We are transitioning through a very large era to the next. China's GDP was report to be better than expected this year that is huge news to the world, specifically for the U.S. it means wen are behind on all things that is "U.S.". our foreign policy is the wrong approach, our relationship between Canada, China, Mexico etc are amongst the worst handled in history technically, Staling knew signing the treaty with Germany was not a guarantee but a time reference for the invasion. Sided and worked with Britain and the U.S.. The shame lies in the h-1 visas and the attack on the education system as a whole. The government has turned their back on the very thing that made the U.S. what it is for many years. Politics are amongst the worst. The responsibilities were in the hands of the U.S. the idol model for the rest of the world to use as inspiration.True Americans are still found just not heard.
Rendered at 22:44:04 GMT+0000 (Coordinated Universal Time) with Vercel.
The answer has nothing to do with rationing. When all software is designed to grab the most attention, serve the most adds, collect the most data, deliver the maximum amount of the cheapest content and generally be as exploitive of its users as possible, no amount of rationing or not rationing will effect it.
McMaster Carr, a company worth a little over a billion, sells metal, bolts, casters, and many other industrial bits and bobs. Their website is infinitely better than Amazons. Why? It has nothing to do with rationing and everything to do with trying to actual deliver a good customer experience.
Two factors:
- middle management do not have contact with either customers or shareholders. They report to managers, and only have managers reporting to them. They're insulated from the actual consequences of their decisions.
- the only way to get recognised for promotion is to create change. Any change. Doesn't matter if it's bad (blame can be shifted), it just needs to be noticed.
If a feature of the product hasn't changed for a long time (because it works for the user), it becomes more of a target for changing. If it changed recently, it creates less impact if it changes again, so isn't a good target. So the product becomes worse over time, not better.
Though, maybe the irony only exists because we use the broad term tech company instead of the more generally accurate terms data harvesting company and attention consuming company.
Their target market is people who need to fix something quickly or get some industrial supply item tomorrow so work isn't stopped.
Massive corporations save money via the end user by taking a chainsaw to customer service. Large tech companies (and many other industries) will just quietly fuck the regular customer over because they can't profitably do customer service at scale. Even Amazon's effectively no questions asked return policy is a band-aid over the forgeries and trash that fill their hyperscale market.
If you have a problem with an item that Mcmaster sends you, they're going to go have a stern talk with the manufacturer. They sell reliability, not race to the bottom capitalism.
(I've also heard that they really take care of their employees. People in jobs that are automated away are always retrained; they'll pay for college with no qualifications, because education is good even if they don't directly benefit, etc.)
I’ve bought a few dozen things from McMaster. For very competent people who know precisely what they want it is good, but I do find I spend an hour looking at the specs of “doodad” which will almost meet my needs when I stumble onto “doodaddy” which is a better fit. After I buy it, a friend points out they sell “doodada” which is even better.
Note I’m an engineer, but didn’t study in North America, so the names of things are different.
McMaster is terrible for product discovery, you have to know the precise name of the thing you want.
For the average person on the street Amazon is better.
It was much the same with search. Google search allowed us to find things faster at first...but over time websites responded to that and started doing SEO crap. Now almost every search results in a billion websites that say the same thing, but shallowly. Yes you can still find some gems but was it worth it?
AI makes things faster, but now it means we can target superificial needs very quickly...and so a new breed of crap software will be built with it. As long as there's a dollar sign at the end of the road.
And of course the number of libraries, known algorithms, etc, is much larger than 8. There's any number of ways to make something new by combining well-understood parts.
I don't think non-expert prompting is pushing the envelope on what can be built because I think the current generation of harnesses and models pushes to using well-understood patterns in well-understood ways but it's certainly moving the line for what has been built. And with more skilled prompting you can build pretty crazy stuff + extensive validation for it.
https://www.science.org/content/article/openai-just-changed-...
Do you think playing dumb is making you look credible or something? The argument isn't "was maths in the dataset" the controversy was if someone with potentially working approaches narrowed the search space enough for the LLM not to fail
> And you think it butresses the argument that the llms can’t do _coding_ tasks they haven’t seen before?
This isn't an argument being made. At least try to respond to what's actually written rather than making yourself look foolish and trying to wordplay your way to a point.
That’s pretty hard to read differently. Can you add the nuance I’m missing?
> By definition an LLM is unable to build what has never been built before because you are going outside the realm of pretraining/RL. You might be able to combine components that already exist in a unique way
It seems you weren’t following that norm and we were talking past each other. That’s fine and I’m happy to leave it at that.
Perhaps in the future you can recognize the local norms before resorting to personal attacks even if those norms are different than the parts of the internet you normally hang out in.
I recall something I once heard about product marketing. The experiment was a product like jam, and they had 3 flavour options available in one case and 12 options available in another. It was a test on how hard it was to make a decision as well as how confident the person was they made the right decision later on.
One of Sartre's points on existentialism was that radical freedom leads to anxiety.
In some sense, AI is removing scarcity. It isn't clear to me that this is the same kind of thing as going from starvation to surplus. It is more like the jam case where we aren't increasing abundance in a total sense, we increasing the options within the existing surplus.
https://en.wikipedia.org/wiki/The_Paradox_of_Choice
I was discussing this with someone else and used the analogy of nuclear energy. Although nuclear plants are highly complex and use forces that have only been understood for the last hundred years, at the end of the day it's still boiling water to create steam to spin a turbine, and nothing about nuclear innovation can engineer away the laws of thermodynamics.
Agents have and will continue to drastically change how we develop code, but it's my belief that drastically changing our software practices to speed up development just means we're going to hit engineering limits faster, rather than being able to overcome those limits entirely.
Unless, or perhaps until, we figure out how to use LLMs effectively at the organization level we're still going to be limited to the "slime mold" level of organizational intelligence.
Will that happen? Most likely, but will it take a generation or ten? I suspect it'll be far harder for society to adapt at that level. We've had Mythical Man Month for decades yet software managers continue repeating the same mistakes.
Because of that the idea of OAI and Anthropic just moving to in-house product development makes sense.
I’m not sure if the job title tho. Software restoration?
The trailblazer was Microsoft Office.
During the hype cycle, making "x for y" is easiest to get funded and exit.
Once the hype crashes and we're in the trough of disillusionment, then people are bearish. So founders either move to the next hype cycle or figure out a more durable business from the ashes
To me, so far it seems like it is and it isn’t simultaneously. Some things are going at insane paces. I see things like reverse engineering and emulation breaking barriers. But we don’t see things like the quality and complexity of highly architected systems getting much better, maybe even getting worse as management pushes velocity onto employees.
I think what this shows is that AI is great at tedium and brute forcing, but it’s not actually good at making substantially better software with less of that “rationed” expertise. At least, not yet.
Tell me I'm wrong.
This is such a weird take, if I need some piece of software why would an ongoing operating cost be having an AI sitting in the critical path.
If the argument is that making software has become cheap because you can use AI to generate software, it doesn't seem to follow that you would build all software as some sort of shell around AI. If I ask claude to build me a task tracking app, I imagine it would build me something with some bog standard data storage and not decide that it needs to keep the tasks in an LLM.
I don't agree with most of this post, but this part stood out as supremely odd. There is a lot of software that isn't made any better by having AI jammed into it as an always on piece of the operating puzzle.
The inversion only applies to products where an AI model sits in the request path, which is a growing share of what’s being built. It definitely doesn't mean all software that will get built (and that wasn't my intent anyway), so maybe my initial sentence in the essay overreaches. I'll make an edit to address this.
But if the answer to "hey Claude, I need to track my gym sessions, what's the best way of doing that?" is "I've built an app for your phone to do that", does that change?
Now that SaaS CRUD apps can be built in 10 minutes, we can maybe do something more worthwhile, no? So much wasted effort has gone into this field already, in my opinion.
It's a different view of scaling. I mean a lot of people have this absurd idea of "super intelligence" which is like the cover of this book
https://en.wikipedia.org/wiki/Supermind_(novel)
i mean what does "intelligence quotient 10,000" even mean? On the other hand it's very easy to imagine a world where we have systems like the best ones we have now but we can really afford to use them.
IQ tests measure, essentially, how fast you can do shallow pattern recognition.
And AI can do these kinds of tests easily.
My IQ is around 120, so it's higher than average, but nothing exceptional. But I'm pretty good at math, so I competed at the national level in my country when I was at school. And as a part of that, I got to attend a summer camp with people who have a REALLY high IQ. Several of them went to win the IMO.
They were able to solve problems much faster than I could during training sessions. Often by the time I started feeling that I had a way to the solution, somebody already had the complete answer ready to be discussed. But with unlimited time to think, I could solve most problems.
In the past, if you wanted to do a task, you had to choose from one of a few (or maybe just one) incumbents and make their tool fit your task. Failing that, build your own tool, but that was expensive. If you did build your own, the tool might have value and could potentially be sold to others.
Going forward, building your own will be much easier and cheaper. Even if the incumbent tools suit the task well, you might choose to build your own for a variety of other reasons. e.g. If the incumbent raises prices or chooses to enshitify. What you will probably not do is adopt some third party's vibe-coded clone of an incumbent's tool, because that won't be customized to your needs and it won't have the track record of the incumbent. You can't trust it or control it so it will be better to build your own.
If an incumbent exists and you can now vibe-code a clone, you probably can't sell that clone "as is" and expect to be successful. You might be able to sell your services to customize and refine tools for small businesses that can't really afford to hire a solitary soul to build their own solution, but don't expect that niche to persist long-term.
There are only so many rust migrations you can do, at some point you gotta ship something user visible which means you have to deal with people which means the code was never the bottleneck even in 2014
There's no reason to think that companies are suddenly going to be spending 1:1 features : quality because both are faster / cheaper .
So expect the low quality to continue, and naturally get worse, because complexity is going up 1000x
In Claude we trust.
2. What is the end goal? The arena looks like that are venture capitalists, Co-founders, founders, angel investors. An MVP that can show off the minimal representation of your vision. At a time where these were steps that each serve their purpose and each come with their own challenges and journeys that are entirely different from one another.
There are oddities that currently exist without little echoed direction, there are two main operating systems for phones and computers. there are only a handful of tech companies with influence and control over direction to which the tech world heads towards. This was more true a few years back, however now we have tech companies directly involved with government and atrocities abroad. The tech companies you remember are no more along with it every piece of software. You might question the exaggeration and dramatized scene before you, but i assure you it is true. OpenAI was a disruptor until 13b was the aid that helped make them pop only after Elon Musk had something to do with the lab before their founded success. It is not that special anymore, just some tech company that comes to disrupt the market for a few years until that main hosting and dependence on the very firms you are looking to challenge. Web3 is not: A blockchain for the internet. A sickening redundant use of AI. Unintentionally reading articles on online of crime and most had section describing the criminal's activity and leading up to the crime with AI prompt they used to convict them. It is clear. We are transitioning through a very large era to the next. China's GDP was report to be better than expected this year that is huge news to the world, specifically for the U.S. it means wen are behind on all things that is "U.S.". our foreign policy is the wrong approach, our relationship between Canada, China, Mexico etc are amongst the worst handled in history technically, Staling knew signing the treaty with Germany was not a guarantee but a time reference for the invasion. Sided and worked with Britain and the U.S.. The shame lies in the h-1 visas and the attack on the education system as a whole. The government has turned their back on the very thing that made the U.S. what it is for many years. Politics are amongst the worst. The responsibilities were in the hands of the U.S. the idol model for the rest of the world to use as inspiration.True Americans are still found just not heard.