Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
nothinkjustai
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
8 ms
·
181.
▲
by
nothinkjustai
7mo ago
No they don’t. What happens is that people want to manipulate the market for a variety of reasons, so they do that, and then really really smart people come along later and try to claim that those manipulators weren’t actually the problem
182.
▲
by
nothinkjustai
7mo ago
Vibe coding requires the sota models to work at all, but someone who knows what they are doing and uses the AI more responsibly can absolutely use the cheaper Chinese models for coding, and they’re often faster too. If I was one of the big
183.
▲
by
nothinkjustai
7mo ago
Beef is yummy, not evil.
184.
▲
by
nothinkjustai
7mo ago
Parlor is so cool, especially since you’re offering it for free. And a great use case for local LLMs.
185.
▲
by
nothinkjustai
7mo ago
> It's widely understood that the big players are making profit on inference. Are they? Or are they just saying that to make their offerings more attractive to investors? Plus I think most people using agents for coding are using su
186.
▲
by
nothinkjustai
7mo ago
What was the metric?
187.
▲
by
nothinkjustai
7mo ago
> Are you also unhappy that in a democracy the 51% choose how the other 49% are going to be governed Yes of course. That is one of it’s fundamental issues.
188.
▲
by
nothinkjustai
7mo ago
I know people say Apple’s commitment to privacy is all talk, and there are valid criticisms of Apple and their business practices, but they seem better than the other big tech companies like Meta, MS, and Google by a very wide margin when i
189.
▲
by
nothinkjustai
7mo ago
Somehow it’ll be Apple
190.
▲
by
nothinkjustai
7mo ago
If results are what matters, why is popular software so buggy and lacking in features?
191.
▲
by
nothinkjustai
7mo ago
Also, wtf are people doing with their lives where a significant amount of time is spent on stuff like this? It only takes a few minutes to book a flight or hotel/airbnb. Shopping for things can be fun, and if it isn’t, again, a few min
192.
▲
by
nothinkjustai
7mo ago
Common sense.
193.
▲
by
nothinkjustai
7mo ago
No
194.
▲
by
nothinkjustai
7mo ago
Wordpress gets you into trouble all the time imo. There are better solutions out there. Every headless cms I’ve used has been much less of a headache compared to Wordpress. I wish I could share your optimism on it, but man, it’s a real pain
195.
▲
by
nothinkjustai
7mo ago
You can even boost its effectiveness by roleplaying with it. I’m not joking. Fully based on vibes, I haven’t done any testing. But it’s part of prompting imo. IMO these things are like a reflection. Present what you want reflected back.
196.
▲
by
nothinkjustai
7mo ago
What is there to anthropodeny?
197.
▲
by
nothinkjustai
7mo ago
I’m guessing you aren’t just asking how an LLM works, but attempting to make the point that humans are also statistical next-token predictors or something? Humans make predictions, that doesn’t mean that’s all we do.
198.
▲
by
nothinkjustai
7mo ago
I don’t see why you can’t use politeness. The thing is a mimic, you “treat” it badly and it mimics how a human might respond. It’s fun to play with, as long as you’re fully cognizant that IT IS NOT A HUMAN
199.
▲
by
nothinkjustai
7mo ago
Idk…my personal experience says it’s probably closer to 100% :)
200.
▲
by
nothinkjustai
7mo ago
It’s a completely different substrate. LLMs don’t have agency, they don’t have a conscious, they don’t have experiences, they don’t learn over time. I’m not saying that the debate is closed, but I also think there is great danger in thinkin
201.
▲
by
nothinkjustai
7mo ago
Oh no. The machine designed to output human-like text is indeed outputting human-like text. I’m half jesting; I think there is a lot of room for debate here, but I also think we shouldn’t anthropomorphize it.
202.
▲
by
nothinkjustai
7mo ago
The copilot plans work with pi and are stupidly cheap for what you get.
203.
▲
by
nothinkjustai
7mo ago
Yeah this is similar to my approach, although with slightly more powerful models. I’m just not having a good time letting the sota models loose on a code base to implement entire features. Spending too much time cleaning up the mess. It’s m
204.
▲
by
nothinkjustai
7mo ago
I don’t understand what he’s actually building with 37k loc/day. In a month that’s 200k loc per project. I tried to go look at his blog but it didn’t load on my phone.