I've tried and tried to get an LLM to delete code, it couldn't do it. I knew a file was 40% bad so what I ended up doing was deleting the file, then asking AI to create the missing file. That's how I got AI to delete code :)
Could it be possible, these firms are optimizing for two things: a) Better performance. b) Gathering data from you to further improve performance later. I've also found the huge amount of planning rather than iteration frustrating. I've felt like I'm teaching a junior!
I think they simply optimize around E2E benchmarks, none of those benchmarks is designed as multi turn assistance to the user, but going from a prompt straight to the final solution.
Exactly. How can "we" develop and encourage benchmarks for multi-turn user assistance?
That is what I want.
I feel like the models and harnesses push much too hard against this workflow -- that they push you towards letting go and vibe coding, with only your discipline (and desire for a quality and maintainable product) holding it back.
I think they are optimizing for one-shot performance because that will drive usage. They can’t afford to look bad in the benchmarks. And if that means consuming an order of magnitude more tokens, well, that’s good for business, too.
I built a passion reference site. A large part of that passion came from knowing and talking to the people I was helping. One person emailing or saying thanks would later help power me through to create more useful articles. Enriching openai/claude/ms/google and no thanks from an individual, has disincentivized me from writing more.
Same here. People knew the website and it was immensely flattering to meet users in the wild. It motivated me to really sweat the small stuff, because people noticed. Now I'm just feeding the slop machine, and it feels pointless.
Is it about the users or the data the users generate. Pretty easy to see the day devs are replaced by the data they themselves generated. Companies are only going to get one chance to grad this data. Similar to the internet cutoff.
Yeah, I meant 'content' in terms of the intrinsic value, the 'nutritional value' underlying the writing... The message, the story, the information content.
Actually it's kind of dystopian to think of it; that the word 'content' has been appropriated to refer to an arbitrarily broad range of media products...
The word 'content' used to be associated with the word 'substance' but in a modern context, it's actually more closely associated with the concept of 'form' as the word emphasizes a variety of media... What happened to the term "multimedia"? IMO this is what I would refer to when some people say content... I mean, there's no content in content... It's empty, it's all smoke and mirrors.
Very correct! Why internal dashboards keep getting rebuild:
https://www.timestored.com/pulse/why-internal-dashboards-get...
It took me a few years to home in on the exact idea you've captured and I work in this exact area. There's a middle layer between UI team and notebook experiments that isn't worth companies building themselves.
The article is ignorant of reality. "The typical under-13 social media user is not a sneaky kid. It’s a family making a decision together. " No, every other kid had it and the parent had no choice else their child would be ostracised. Their example of kids learning about volcanoes in youtube. Ha! Go look at the view number for mindless nonsense... Minecraft blabbering then find me a volcano with more child views.
I wouldn’t mind YouTube if my kids asked to watch videos of volcanoes.
Instead they always end up trying to watch the most annoying people possible playing video games for a few minutes before we ask them to switch it off.
If YouTube really cared about kids they should allow users to pick a subset of channels and never ever mention again anything else.
Alas, we all know this will never happen because YouTube doesn’t care about what people watch as long as they come back watching some more