There is hype around AI and agentic dev. It's not subjective. My opinion is that it's a valid factor to consider if you evaluate tech. We've seen with i.e. microservices that people followed the hype because it seemed sound, but it was eventually terrible for many. Not everything is for everyone just because it's there.
Can people explore things they care about? Yes. Can I have an opinion about what they do if they make a public post about it. Yes.
> Most of it reads like too high expectation of overhyped technology.
I don’t see how this is true. By what metric might “most” lead to a true statement?
The post expresses a very low amount of hype, if any. It predominantly an account of trying to make agents work.
The above comment also seems dismissive. I suggest giving the author some credit. The article is much more valuable than most of what I read on AI from various pundits who seem to repeat the same narrative. Thankfully, this article teaches without preaching or positing some predefined narrative.
“Most” in English means more than half. And there is an underlying meaning, stated or implied, by which we can figure out if we are in the “most” category. So, to restate, by what meaning is >50% of the article the result of too high expectations? Please show us some detail; don’t just tell us that it is that way.
The best I can puzzle together is that you are suggesting “there is AI hype” and (hand waving) this undermines the article (somehow). I don’t see it. You don’t explain specific weaknesses about the article itself other than it is less detailed that you would like. You said you demand depth, so it would seem fair to respond with at least as much detail as the author provided.
We are at the point where saying “AI is overhyped” is very overloaded, making its information value low or even negative. Instead, I suggest being specific. For example, if you think OpenAI has made [specific] investments or product choices that are unlikely to materialize over a quarter or a year, say that and explain why. My hope here is to learn and reduce the chances of talking past another person.
Even if I were to use the vague language of saying “we are in a hype bubble” (which I don’t because I think it is not useful nor clear), most examples of hype are associated with important and substantive changes or lessons to learn.
So when an author writes up their lessons learned about making agents work better, I find it a bit low-effort to see someone else dismiss that article without any detail at all.
To move this forward, what do you think the author got wrong about agentic AI, specifically?
There is hype around AI and agentic dev. It's not subjective. My opinion is that it's a valid factor to consider if you evaluate tech. We've seen with i.e. microservices that people followed the hype because it seemed sound, but it was eventually terrible for many. Not everything is for everyone just because it's there.
Can people explore things they care about? Yes. Can I have an opinion about what they do if they make a public post about it. Yes.