5 ms·

I'm tired of constantly debating the same thing again and again. Where are the products? Where is some great performing software all LLM/agent crafted? All I se

I'm tired of constantly debating the same thing again and again. Where are the products? Where is some great performing software all LLM/agent crafted? All I see is software bloatness and decline. Where is Discord that uses just a bunch of hundreds megs of ram? Where is unbloated faster Slack? Where is the Excel killer? Fast mobile apps? Browsers and the web platform improved? Why Cursor team don't use Cursor to get rid of vscode base and code its super duper code editor? I see tons of talking and almost zero products.

krageon 10mo ago
https://www.anthropic.com/research/how-ai-is-transforming-work-at-anthropic https://www.anthropic.com/research/how-ai-is-transforming-wo... see "How much work can be fully delegated to Claude?": "Although engineers use Claude frequently, more than half said they can “fully delegate” only between 0-20% of their work to Claude" There won't be anything like you're asking for, even the vendors themselves (they'll be the most positive and most enthousiastic about using it) can't do this with them.
- hollowturtle 10mo ago
  I'm not asking for it, i'm asking to stop bulshitting about ai
  - krageon 10mo ago
    My point is that you can ignore every article about ai being super good as long as you see the vendor research (that you read once a year or less) is still the same. It saves everyone a lot of frustration. As for why it keeps appearing here, people like being excited. It's not about the truth, so asking for it is missing the point.
    - hollowturtle 10mo ago
      I agree partially, my main frustration comes from "network effects" of people reading these statements without taking them with a grain of salt
- D-Machine 10mo ago
  Thank you for linking this very useful and much more realistic / grounded stat.
konhi 10mo ago
I share similar feelings. I feel like I'm reading the same comments about LLM since a year, only model version changes. Obviously there's improvement in the models and tooling, but the debate seems very artificial.
michalsustr 10mo ago
Hi, I’m building such one: https://minfx.ai/ https://minfx.ai/ Still early, but iterating really fast!
lunias 10mo ago
This is true. I think most people are mostly using AI at work to fix bugs in existing codebases. A smaller group of people are benchmarking AI by giving it ideas for apps that no one needs and seeing if it can get close. The smallest group of people is actually designing new software and asking the AI to iterate on it.
g947o 10mo ago
This deserves more upvotes. Even if there is a "fully vibe-coded" product that has real customers, the fact that it's vibe-coded means that others can do the same. Unless you have a secret LLM or some magical prompts that make the code better/more efficient than your competitions, your vibe coded product has no advantage over competition and no moat. What actually matters is everything else -- user experience (which requires hours of meetings and usability studies), integration with own/other people's products, business, marketing, sales etc, much of which you can't vibe code your way to success.
- enraged_camel 10mo ago
  >> Even if there is a "fully vibe-coded" product that has real customers, the fact that it's vibe-coded means that others can do the same. But that's precisely why you don't hear about these products: the creators don't disclose that they were vibe-coded, because if they do, that invites competition. I personally know of four vibe-coded products that generate over $10k/mo. Two of them were made by one friend, one was made by another, and the last one by my cousin. None of these people are developers. But they are making real money.
  - hollowturtle 10mo ago
    We all have a cousin that makes $10k/mo and has super powers
    - enraged_camel 10mo ago
      My cousin definitely does not have superpowers. ;)
    - jayrot 10mo ago
      You wouldn't know her, though. She lives in Canada.
    - Rapzid 10mo ago
      Website roulette probably has a 50% shot at loading a blog written by a digital nomad who makes a living off some SEO side project that pays for their Asia-Pacific island lifestyle... All pre-LLM.
jwblackwell 10mo ago
With all due respect somebody could launch a version of Discord that's 10x faster tomorrow and nobody would know about it It's very difficult to unseat those incumbents, especially those with strong network effects. Plus the people that work in those larger companies are not at the edge of AI coding at all and not motivated to rock the boat
- hollowturtle 10mo ago
  Totally disagree. One example is Zed which is very well known and it's faster than any other editor, wasn't built with AI though. > People on larger companies are not at the edge of AI coding False Microsoft is all in with Copilot, and I can't believe the company that created Copilot doesn't use it internally, I'd rather say they should be the ones that would know how to master it! Yet no better vscode, still bloated teams etc etc
  - PretzelPirate 10mo ago
    > Yet no better vscode, still bloated teams etc etc Why do you assume that Microsoft would focus on building a better (to you) VSCode or less bloated Teams? I assume they'd use Github Copilot to make a more profitable VSCode and Teams, which doesn't require focusing on speed and bloat.
    - hollowturtle 10mo ago
      I'm not assuming, the whole narrative goes like "software development is a solved problem and a sunken cost" Ok if the cost is low why not then? It makes sense to improve your product and strenghten your market position
      - PretzelPirate 10mo ago
        > Ok if the cost is low why not then? There are still budgets that constrain on how much you can achieve. If Microsoft thinks using AI to add more AI features will be more profitable than increasing performance, the that's where they'll spend their budget.
weitendorf 10mo ago
Except for maybe an "Excel killer", all those things you listed are not things people are willing to pay for. Also agents are bad at that kind of work (most devs are bad at that stuff, it's why it was something people whined about even before agents). And funnily enough there are products and tools that are essentially less bloated slack/discord. Have you heard of https://stoat.chat/ https://stoat.chat/ (aka revolt) or https://pumble.com/ https://pumble.com/ or https://meet.jit.si/ https://meet.jit.si/? If not I would guess it's for one of two reasons: not caring enough about these problems to even go looking for them yourself, or their lack of "bloatedness" resulting in them not being a mature/fully featured enough product to be worth marketing or adopting. If you'd like to see a product mostly made with agents/for agents you can check out mine at https://statue.dev/ https://statue.dev/ - we're making a static site generator with a templating and component system paired with user-story driven "agentic workflows" (~blueprints/playbooks for common user actions like "I need to add a new page and list it on the navbar" or "create a site from the developer portfolio template personalized for my github"). I would guess most other projects are probably in a similar situation as we are: agentic developer tools have only really been good enough to heavily use/build products around for a few months, so it's a typical few-month-old project. But agents definitely made it easier to build.
- hollowturtle 10mo ago
  Not willing to pay for? How can you be sure? For example explain then why many gamers are ditching Windows for Linux and buying hardware from Valve... There must be a reason. Every person I talked to that uses Excel hate how slow it is, same for teams and many other products. Finally, were the mentioned products built with vibe coding?
  - weitendorf 10mo ago
    Generally if something is fast enough/efficient enough that a paying customer can use it without having to worry or actively think about performance and un-bloatedness, that's enough for them. The only people who might complain still are developers who are bothered by the inefficiency and are technically literate enough to notice it, and maybe the users with less powerful/capable devices than the ones the big paying customers use. Generally these groups of people are not the actual customers of these products. The people who actually pay for slack and discord (eg enterprises that need workplace chat app and decided to go with the "gold standard", consumers with discord servers and such) need the features/tradeoffs choosing featuers over efficiency causing that bloat. They just don't all need the exact same set of those features as the other customers. So because customers are willing to pay for all these features the product tries to ship all of them and becomes bloated. > Every person I talked to that uses Excel hate how slow it is But do they make the purchasing decisions behind using Excel? To be clear I am not really arguing that bloat/overly enterprisey products are good. What I mean that you don't see the world exploding with more elegant products now with agents for the same reason you didn't see the world exploding with them before agents either: the people who pay for those products and build them for a living are not incentivized or necessarily even rewarded for choosing to make them more efficient or elegant when there are other things that customers are asking for with more $$$ behind them.
broken_ceiling 10mo ago
This argument falls a little flat when you consider how much software may or may not be written inside one's own personal work flow, or to scale that up, inside a small business. The idea that a small business doing >1mil revenue can now hire a dev or two, and build out a fairly functional domain-driven system should not be understated. The democratization of software, and the lowering of the barriers to entry to basic CRUD apps, may not necessarily show up in a TAM report... Do you need a killer app that treads into unicorn territory to prove it's impact? What about a million apps that displace said unicorn potentials by removing the need for a COTS? Oh, and remember, the iPhone was revolutionary but it was diffused so slowly into the greater economy, the impact on global GDP was basically negligent. Actually, almost all the perceived grandiose tech jumps did not magically produce huge GDP gains overnight.
- hollowturtle 10mo ago
  Your argument falls a little flat considering that you mention "hire a dev or two" while the whole narrative is "we don't need software engineers anymore" and Anthropic alone declares that "Although engineers use Claude frequently, more than half said they can “fully delegate” only between 0-20% of their work to Claude" https://www.anthropic.com/research/how-ai-is-transforming-work-at-anthropic https://www.anthropic.com/research/how-ai-is-transforming-wo...
  - broken_ceiling 10mo ago
    When was I arguing about job displacement or the replacing of engineers? You are projecting hard, and reaching. If anything, I am in the camp that accessibility to custom tooling equals a net positive of devs down the line. In the short term, it may be a bumpy road as the tools progress (even if incrementally), but my long term take is that you may see engineering teams blossom in smaller market operations. When it comes to objectivity, people with your line of thinking is what I try to avoid, as it is clear you feel threatened by the progress of coding tools. That link doesn't change much about what I said, or for that matter, what you said. You were commenting on the lack of a killer app, and I just said it may be diffusing slowly in different ways. You are fixating on the "whole narrative" because you feel threatened - rightfully so, but again, that type of hyperbole doesn't belong in a constructive and grounded conversation about the impact AI may or may not have.
quest88 10mo ago
Anecdotally I had Gemini convert a simple react native app to swift in two prompts. If it's that simple then maybe we will see less of the chromium desktop apps
- hollowturtle 10mo ago
  I'd argue the contrary, YOU KNOW you have the option, ease of entering doesn't mean they will know how to choose better, they will just vibe code more electron apps. In fact my prediction is not there will be less Electron apps but more
taytus 10mo ago
Could someone explain this to me? I have the same question: why Cursor team don't use Cursor to get rid of vscode base and code its super duper code editor?
ebiester 10mo ago
The bigger the product, the harder this is. However, I think the biggest thing is the replacement of products. We are in a place where he talked about replacing two products his wife was using with custom software. I personally have used LLMs to build things that are valuable for me that I just don't have time for otherwise.
maxdo 10mo ago
who told you that mb of ram is a definition of success? Opus was out only few months, and it will take time to get this new wave to market. i can assure you my team become way more productive because of opus. not a single developer but an etnire team.
- hollowturtle 10mo ago
  It's a definition of what runs and what not on consumer grade computers, Discord has a routine that now checks if memory goes over a certain threshold and eventually restart itselfs, this is a measure of engineering total failure imo
xnx 10mo ago
AI amplified development has the most impact on build-vs-buy decisions. We should expect the decreased difficulty of creating software to drive down prices.
- hollowturtle 10mo ago
  > decreased difficulty of creating software to drive down prices. And here we go again, if difficulty has been decreased so much, where are the fixes or the products?

krageon 10mo ago

https://www.anthropic.com/research/how-ai-is-transforming-work-at-anthropic https://www.anthropic.com/research/how-ai-is-transforming-wo... see "How much work can be fully delegated to Claude?": "Although engineers use Claude frequently, more than half said they can “fully delegate” only between 0-20% of their work to Claude" There won't be anything like you're asking for, even the vendors themselves (they'll be the most positive and most enthousiastic about using it) can't do this with them.

hollowturtle 10mo ago
I'm not asking for it, i'm asking to stop bulshitting about ai
- krageon 10mo ago
  My point is that you can ignore every article about ai being super good as long as you see the vendor research (that you read once a year or less) is still the same. It saves everyone a lot of frustration. As for why it keeps appearing here, people like being excited. It's not about the truth, so asking for it is missing the point.
  - hollowturtle 10mo ago
    I agree partially, my main frustration comes from "network effects" of people reading these statements without taking them with a grain of salt
D-Machine 10mo ago
Thank you for linking this very useful and much more realistic / grounded stat.

konhi 10mo ago

I share similar feelings. I feel like I'm reading the same comments about LLM since a year, only model version changes. Obviously there's improvement in the models and tooling, but the debate seems very artificial.

michalsustr 10mo ago

Hi, I’m building such one: https://minfx.ai/ https://minfx.ai/ Still early, but iterating really fast!

lunias 10mo ago

This is true. I think most people are mostly using AI at work to fix bugs in existing codebases. A smaller group of people are benchmarking AI by giving it ideas for apps that no one needs and seeing if it can get close. The smallest group of people is actually designing new software and asking the AI to iterate on it.

g947o 10mo ago

This deserves more upvotes. Even if there is a "fully vibe-coded" product that has real customers, the fact that it's vibe-coded means that others can do the same. Unless you have a secret LLM or some magical prompts that make the code better/more efficient than your competitions, your vibe coded product has no advantage over competition and no moat. What actually matters is everything else -- user experience (which requires hours of meetings and usability studies), integration with own/other people's products, business, marketing, sales etc, much of which you can't vibe code your way to success.

enraged_camel 10mo ago
>> Even if there is a "fully vibe-coded" product that has real customers, the fact that it's vibe-coded means that others can do the same. But that's precisely why you don't hear about these products: the creators don't disclose that they were vibe-coded, because if they do, that invites competition. I personally know of four vibe-coded products that generate over $10k/mo. Two of them were made by one friend, one was made by another, and the last one by my cousin. None of these people are developers. But they are making real money.
- hollowturtle 10mo ago
  We all have a cousin that makes $10k/mo and has super powers
  - enraged_camel 10mo ago
    My cousin definitely does not have superpowers. ;)
  - jayrot 10mo ago
    You wouldn't know her, though. She lives in Canada.
  - Rapzid 10mo ago
    Website roulette probably has a 50% shot at loading a blog written by a digital nomad who makes a living off some SEO side project that pays for their Asia-Pacific island lifestyle... All pre-LLM.

jwblackwell 10mo ago

With all due respect somebody could launch a version of Discord that's 10x faster tomorrow and nobody would know about it It's very difficult to unseat those incumbents, especially those with strong network effects. Plus the people that work in those larger companies are not at the edge of AI coding at all and not motivated to rock the boat

hollowturtle 10mo ago
Totally disagree. One example is Zed which is very well known and it's faster than any other editor, wasn't built with AI though. > People on larger companies are not at the edge of AI coding False Microsoft is all in with Copilot, and I can't believe the company that created Copilot doesn't use it internally, I'd rather say they should be the ones that would know how to master it! Yet no better vscode, still bloated teams etc etc
- PretzelPirate 10mo ago
  > Yet no better vscode, still bloated teams etc etc Why do you assume that Microsoft would focus on building a better (to you) VSCode or less bloated Teams? I assume they'd use Github Copilot to make a more profitable VSCode and Teams, which doesn't require focusing on speed and bloat.
  - hollowturtle 10mo ago
    I'm not assuming, the whole narrative goes like "software development is a solved problem and a sunken cost" Ok if the cost is low why not then? It makes sense to improve your product and strenghten your market position
    - PretzelPirate 10mo ago
      > Ok if the cost is low why not then? There are still budgets that constrain on how much you can achieve. If Microsoft thinks using AI to add more AI features will be more profitable than increasing performance, the that's where they'll spend their budget.

weitendorf 10mo ago

Except for maybe an "Excel killer", all those things you listed are not things people are willing to pay for. Also agents are bad at that kind of work (most devs are bad at that stuff, it's why it was something people whined about even before agents). And funnily enough there are products and tools that are essentially less bloated slack/discord. Have you heard of https://stoat.chat/ https://stoat.chat/ (aka revolt) or https://pumble.com/ https://pumble.com/ or https://meet.jit.si/ https://meet.jit.si/? If not I would guess it's for one of two reasons: not caring enough about these problems to even go looking for them yourself, or their lack of "bloatedness" resulting in them not being a mature/fully featured enough product to be worth marketing or adopting. If you'd like to see a product mostly made with agents/for agents you can check out mine at https://statue.dev/ https://statue.dev/ - we're making a static site generator with a templating and component system paired with user-story driven "agentic workflows" (~blueprints/playbooks for common user actions like "I need to add a new page and list it on the navbar" or "create a site from the developer portfolio template personalized for my github"). I would guess most other projects are probably in a similar situation as we are: agentic developer tools have only really been good enough to heavily use/build products around for a few months, so it's a typical few-month-old project. But agents definitely made it easier to build.

hollowturtle 10mo ago
Not willing to pay for? How can you be sure? For example explain then why many gamers are ditching Windows for Linux and buying hardware from Valve... There must be a reason. Every person I talked to that uses Excel hate how slow it is, same for teams and many other products. Finally, were the mentioned products built with vibe coding?
- weitendorf 10mo ago
  Generally if something is fast enough/efficient enough that a paying customer can use it without having to worry or actively think about performance and un-bloatedness, that's enough for them. The only people who might complain still are developers who are bothered by the inefficiency and are technically literate enough to notice it, and maybe the users with less powerful/capable devices than the ones the big paying customers use. Generally these groups of people are not the actual customers of these products. The people who actually pay for slack and discord (eg enterprises that need workplace chat app and decided to go with the "gold standard", consumers with discord servers and such) need the features/tradeoffs choosing featuers over efficiency causing that bloat. They just don't all need the exact same set of those features as the other customers. So because customers are willing to pay for all these features the product tries to ship all of them and becomes bloated. > Every person I talked to that uses Excel hate how slow it is But do they make the purchasing decisions behind using Excel? To be clear I am not really arguing that bloat/overly enterprisey products are good. What I mean that you don't see the world exploding with more elegant products now with agents for the same reason you didn't see the world exploding with them before agents either: the people who pay for those products and build them for a living are not incentivized or necessarily even rewarded for choosing to make them more efficient or elegant when there are other things that customers are asking for with more $$$ behind them.

broken_ceiling 10mo ago

This argument falls a little flat when you consider how much software may or may not be written inside one's own personal work flow, or to scale that up, inside a small business. The idea that a small business doing >1mil revenue can now hire a dev or two, and build out a fairly functional domain-driven system should not be understated. The democratization of software, and the lowering of the barriers to entry to basic CRUD apps, may not necessarily show up in a TAM report... Do you need a killer app that treads into unicorn territory to prove it's impact? What about a million apps that displace said unicorn potentials by removing the need for a COTS? Oh, and remember, the iPhone was revolutionary but it was diffused so slowly into the greater economy, the impact on global GDP was basically negligent. Actually, almost all the perceived grandiose tech jumps did not magically produce huge GDP gains overnight.

hollowturtle 10mo ago
Your argument falls a little flat considering that you mention "hire a dev or two" while the whole narrative is "we don't need software engineers anymore" and Anthropic alone declares that "Although engineers use Claude frequently, more than half said they can “fully delegate” only between 0-20% of their work to Claude" https://www.anthropic.com/research/how-ai-is-transforming-work-at-anthropic https://www.anthropic.com/research/how-ai-is-transforming-wo...
- broken_ceiling 10mo ago
  When was I arguing about job displacement or the replacing of engineers? You are projecting hard, and reaching. If anything, I am in the camp that accessibility to custom tooling equals a net positive of devs down the line. In the short term, it may be a bumpy road as the tools progress (even if incrementally), but my long term take is that you may see engineering teams blossom in smaller market operations. When it comes to objectivity, people with your line of thinking is what I try to avoid, as it is clear you feel threatened by the progress of coding tools. That link doesn't change much about what I said, or for that matter, what you said. You were commenting on the lack of a killer app, and I just said it may be diffusing slowly in different ways. You are fixating on the "whole narrative" because you feel threatened - rightfully so, but again, that type of hyperbole doesn't belong in a constructive and grounded conversation about the impact AI may or may not have.

quest88 10mo ago

Anecdotally I had Gemini convert a simple react native app to swift in two prompts. If it's that simple then maybe we will see less of the chromium desktop apps

hollowturtle 10mo ago
I'd argue the contrary, YOU KNOW you have the option, ease of entering doesn't mean they will know how to choose better, they will just vibe code more electron apps. In fact my prediction is not there will be less Electron apps but more

taytus 10mo ago

Could someone explain this to me? I have the same question: why Cursor team don't use Cursor to get rid of vscode base and code its super duper code editor?

ebiester 10mo ago

The bigger the product, the harder this is. However, I think the biggest thing is the replacement of products. We are in a place where he talked about replacing two products his wife was using with custom software. I personally have used LLMs to build things that are valuable for me that I just don't have time for otherwise.

maxdo 10mo ago

who told you that mb of ram is a definition of success? Opus was out only few months, and it will take time to get this new wave to market. i can assure you my team become way more productive because of opus. not a single developer but an etnire team.

hollowturtle 10mo ago
It's a definition of what runs and what not on consumer grade computers, Discord has a routine that now checks if memory goes over a certain threshold and eventually restart itselfs, this is a measure of engineering total failure imo

xnx 10mo ago

AI amplified development has the most impact on build-vs-buy decisions. We should expect the decreased difficulty of creating software to drive down prices.

hollowturtle 10mo ago
> decreased difficulty of creating software to drive down prices. And here we go again, if difficulty has been decreased so much, where are the fixes or the products?