4 ms·

Honestly? You're thinking about it wrong. AI can make most products more useful. That doesn't mean it's going to always be implemented well. Think about all th

Honestly? You're thinking about it wrong. AI can make most products more useful. That doesn't mean it's going to always be implemented well. Think about all the companies that were selling boxed desktop software in the 90s. The ones that survived almost universally found some way to incorporate networking/internet into their product. At the time, there was plenty of skepticism, and you could have said they were chasing a fad. But to early adopters of the internet, the value was plainly obvious. Of course, many tried to adapt but executed poorly, so they still didn't make it through the paradigm shift. That doesn't mean ignoring the shift was a better plan.

sumeno 1y ago
> AI can make most products more useful That's what the people shoving it down our throats keep saying but the reality is the vast majority of products don't need AI and are often made worse by it.
- danenania 1y ago
  > are often made worse by it No disagreement on this, but it's an execution problem.
  - throwaway314155 1y ago
    When the execution problem is as ubiquitous as it is, you just call it a problem.
  - stonogo 1y ago
    I've been hearing that about communism for a long time too
    - baobun 1y ago
      That's true though. Just as democratic as the DPRK, as united as the USA and as green as Greenland...
  - xvello 1y ago
    This is the fallacy that LLM-boosters are spreading: "AI can never fail, it can only be failed". But when the fail ratio is so high, it cannot be only an execution problem. It is proof that the LLM technology is over-hyped way over the small niche where it's cost/benefit ratio can be positive. https://edition.cnn.com/2025/03/27/tech/apple-ai-artificial-intelligence/index.html https://edition.cnn.com/2025/03/27/tech/apple-ai-artificial-...
    - danenania 1y ago
      Many hard things have a high rate of execution failure.
listeria 1y ago
Shouldn't you have added a disclaimer, Mr. Founder of Plandex: an open source AI coding agent?
- danenania 1y ago
  What should my disclaimer be? That I'm building something which is open source and consistent with my beliefs?
  - listeria 1y ago
    I was actually thinking about something like this: "disclaimer: I'm the founder of Plandex: an open source AI coding agent" Note that I wouldn't personally add "open source", I don't think it adds to the discussion, but it's what I would have expected to see.
    - danenania 1y ago
      Is it relevant? This is not a post about AI coding, and Plandex is not a startup that "went from X to AI-powered X". It also has very little to do with adding AI features to a product... it's a general purpose coding agent, not a tool for building AI features specifically. I mean, I discuss AI a lot on HN. Do I need to include a disclaimer on every comment?
  - throwaway314155 1y ago
    I think perhaps more important is wether or not you've accepted funding from venture capital, as this would substantially reduce the moral argument made by claiming you're making something open source (as OSS essentially the playbook for developing initial growth in VC land).
    - danenania 1y ago
      I've raised a small amount of funding yes. The project wouldn't exist otherwise. I have a family to feed. But the open source project is free, full-featured, MIT licensed, and will stay that way. It also has paying customers via cloud-hosting options, and the economics are working out well so far, fwiw. Lots of work to do, but it's on track to be both a sustainable open source project and a sustainable business. I'm sorry if that still doesn't meet your lofty moral standards—all I can really say is I'm trying my best to get the balance right.
Retric 1y ago
Future AI possibly, but what’s currently available needs to actually be fit for purpose or it’s just a waste of resources. Many 90’s games added video elements because that’s ‘something new we can do’ without actually improving the gameplay. It’s that kind of low effort tacking on of cool features that you really need to avoid.
bisby 1y ago
There is also the caveat that some companies tried things that work now, but didn't work back then. I don't remember the specific company, but something like Uber was tried back then. and it just didn't work. "Internet is available" is not the same as "internet is ubiquitous" and some ideas require the ubiquity for it to work out, even if the execution was otherwise fine. "LLMs are pretty neat" is not the same as "AGI is ubiquitous." So there are some AI products that people will try, that will fail horrendously, and it won't be because it's a bad idea or executed poorly, but it's because the AI behind the idea isn't ready yet.
- danenania 1y ago
  Yeah that's true, but I would still categorize it as an execution problem. If your product (or your AI feature) doesn't work because the models aren't ready yet, it means you failed at design and implementation imho. It doesn't mean there's no plausible AI integration that would make the product more valuable. It's like if you choose the wrong database and it makes your product slow. It probably doesn't imply you shouldn't be using any database; you just need a different one.
cj 1y ago
I don't think the GP meant to say that AI doesn't make products more useful. I read it as them being frustrated with the over-use of "AI-powered X" -- when everything is AI powered, is anything AI powered? Favorite quote from Garry Tan: If everything is bold, nothing is bold! "AI-powered" as a descriptor is incredibly overused to the point where it no longer elicits the same excitement that it used to 1-2 years ago. I think everyone agrees that AI incorporated into a product can definitely make the product more useful. But let's stop describing our products as "AI-powered" and replace it with a descriptor that actually communicates the value the product is providing. Marketing for AI products would be more successful IMO if they stop using the words "AI-powered" and "LLM" and "agents" and "tokens", etc... users are overwhelmed with AI overload. Throw away the buzz words and use messaging that communicates the specific value that the AI adds. I think the days of consumers seeking out "AI" for the sake of AI is coming to an end. Consumers want to know what the AI will do for them. They don't care that it's AI-powered.
- danenania 1y ago
  That’s fair. I think the OP put it in a much more dismissive way than you did though. If you’re building a product and it’s not working, pivoting in the direction of AI could definitely be rational—not because of the hype, or at least not only because of that (it doesn’t hurt). But because it could give you a better chance of building a more useful and differentiated product. Implying that any entrepreneur who does this is just following the herd is pretty uncharitable imo. Again, it would be like saying that about people pivoting to web in the early 2000s or mobile in the 2010s. Not every trend is fake.

sumeno 1y ago

> AI can make most products more useful That's what the people shoving it down our throats keep saying but the reality is the vast majority of products don't need AI and are often made worse by it.

danenania 1y ago
> are often made worse by it No disagreement on this, but it's an execution problem.
- throwaway314155 1y ago
  When the execution problem is as ubiquitous as it is, you just call it a problem.
- stonogo 1y ago
  I've been hearing that about communism for a long time too
  - baobun 1y ago
    That's true though. Just as democratic as the DPRK, as united as the USA and as green as Greenland...
- xvello 1y ago
  This is the fallacy that LLM-boosters are spreading: "AI can never fail, it can only be failed". But when the fail ratio is so high, it cannot be only an execution problem. It is proof that the LLM technology is over-hyped way over the small niche where it's cost/benefit ratio can be positive. https://edition.cnn.com/2025/03/27/tech/apple-ai-artificial-intelligence/index.html https://edition.cnn.com/2025/03/27/tech/apple-ai-artificial-...
  - danenania 1y ago
    Many hard things have a high rate of execution failure.

listeria 1y ago

Shouldn't you have added a disclaimer, Mr. Founder of Plandex: an open source AI coding agent?

danenania 1y ago
What should my disclaimer be? That I'm building something which is open source and consistent with my beliefs?
- listeria 1y ago
  I was actually thinking about something like this: "disclaimer: I'm the founder of Plandex: an open source AI coding agent" Note that I wouldn't personally add "open source", I don't think it adds to the discussion, but it's what I would have expected to see.
  - danenania 1y ago
    Is it relevant? This is not a post about AI coding, and Plandex is not a startup that "went from X to AI-powered X". It also has very little to do with adding AI features to a product... it's a general purpose coding agent, not a tool for building AI features specifically. I mean, I discuss AI a lot on HN. Do I need to include a disclaimer on every comment?
- throwaway314155 1y ago
  I think perhaps more important is wether or not you've accepted funding from venture capital, as this would substantially reduce the moral argument made by claiming you're making something open source (as OSS essentially the playbook for developing initial growth in VC land).
  - danenania 1y ago
    I've raised a small amount of funding yes. The project wouldn't exist otherwise. I have a family to feed. But the open source project is free, full-featured, MIT licensed, and will stay that way. It also has paying customers via cloud-hosting options, and the economics are working out well so far, fwiw. Lots of work to do, but it's on track to be both a sustainable open source project and a sustainable business. I'm sorry if that still doesn't meet your lofty moral standards—all I can really say is I'm trying my best to get the balance right.

Retric 1y ago

Future AI possibly, but what’s currently available needs to actually be fit for purpose or it’s just a waste of resources. Many 90’s games added video elements because that’s ‘something new we can do’ without actually improving the gameplay. It’s that kind of low effort tacking on of cool features that you really need to avoid.

bisby 1y ago

There is also the caveat that some companies tried things that work now, but didn't work back then. I don't remember the specific company, but something like Uber was tried back then. and it just didn't work. "Internet is available" is not the same as "internet is ubiquitous" and some ideas require the ubiquity for it to work out, even if the execution was otherwise fine. "LLMs are pretty neat" is not the same as "AGI is ubiquitous." So there are some AI products that people will try, that will fail horrendously, and it won't be because it's a bad idea or executed poorly, but it's because the AI behind the idea isn't ready yet.

danenania 1y ago
Yeah that's true, but I would still categorize it as an execution problem. If your product (or your AI feature) doesn't work because the models aren't ready yet, it means you failed at design and implementation imho. It doesn't mean there's no plausible AI integration that would make the product more valuable. It's like if you choose the wrong database and it makes your product slow. It probably doesn't imply you shouldn't be using any database; you just need a different one.

cj 1y ago

I don't think the GP meant to say that AI doesn't make products more useful. I read it as them being frustrated with the over-use of "AI-powered X" -- when everything is AI powered, is anything AI powered? Favorite quote from Garry Tan: If everything is bold, nothing is bold! "AI-powered" as a descriptor is incredibly overused to the point where it no longer elicits the same excitement that it used to 1-2 years ago. I think everyone agrees that AI incorporated into a product can definitely make the product more useful. But let's stop describing our products as "AI-powered" and replace it with a descriptor that actually communicates the value the product is providing. Marketing for AI products would be more successful IMO if they stop using the words "AI-powered" and "LLM" and "agents" and "tokens", etc... users are overwhelmed with AI overload. Throw away the buzz words and use messaging that communicates the specific value that the AI adds. I think the days of consumers seeking out "AI" for the sake of AI is coming to an end. Consumers want to know what the AI will do for them. They don't care that it's AI-powered.

danenania 1y ago
That’s fair. I think the OP put it in a much more dismissive way than you did though. If you’re building a product and it’s not working, pivoting in the direction of AI could definitely be rational—not because of the hype, or at least not only because of that (it doesn’t hurt). But because it could give you a better chance of building a more useful and differentiated product. Implying that any entrepreneur who does this is just following the herd is pretty uncharitable imo. Again, it would be like saying that about people pivoting to web in the early 2000s or mobile in the 2010s. Not every trend is fake.