Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
jasfi
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
9 ms
·
91.
▲
by
jasfi
2y ago
Does this explain some LLM behavior when they refuse to give up their position? I've heard about this and I've seen it myself. Not saying it's because of the architecture, but perhaps because of human conversations in the tra
92.
▲
by
jasfi
2y ago
Not op, but ColorNote is good.
93.
▲
by
jasfi
2y ago
I'm building something similar, with a focus on quality. There's a wait-list you can sign up for in the meantime: https://aiconstrux.com
94.
▲
by
jasfi
2y ago
Probably, without any sources or anything related online it seems to be fiction.
95.
▲
by
jasfi
2y ago
This could be at the core of "do things that don't scale".
96.
▲
by
jasfi
2y ago
Two things: - An AI web app builder that aims to overcome many of the problems that similar solutions have. I have a wait-list you can sign-up for while I build the prototype: https://aiconstrux.com - A crypto market indicator:
97.
▲
by
jasfi
2y ago
The Civ series does the same thing for me. I've also moved on from games, although I think from time to time they may be useful for stimulating creativity.
98.
▲
by
jasfi
2y ago
The headline is actually misleading (from the original article). The first paragraph clarifies that they're talking about the overconsumption of read meat.
99.
▲
by
jasfi
2y ago
I'd go with that option too. I don't think users want to be surprised with being experimented on. Some users could take it worse than others.
100.
▲
by
jasfi
2y ago
Moving fast is often crucial, but you typically also get more readable code, which is also worth it. A query/DML that is badly optimized is usually very complex, which means that the ORM's syntax isn't a good fit. No problem,
101.
▲
by
jasfi
2y ago
AGI would look like that, given enough time. It would be somewhat plausible for an AGI civilization to go off on their own, becoming completely separate from humanity. They'd have a much better chance at becoming a spacefaring civ.
102.
▲
by
jasfi
2y ago
Coinbase is probably an example.
103.
▲
by
jasfi
2y ago
I think you should work with your doctor on resolving this. Using alcohol as a crutch sounds like it could lead to more problems. Are you on any meds? Is this not perhaps a side-effect of something? Have you tried Ashwagandha? Perhaps ask y
104.
▲
by
jasfi
2y ago
They create variants from a single video from a real creator. They don't say if the examples on the career page are AI generated variants, but if they are, they're not generated from scratch.
105.
▲
by
jasfi
2y ago
A lot of business people want to get something functional that they can sell, and hire a programmer if/when they can afford one. That niche is seeing a lot of uptake with regards to LLM based approaches. This works for them because an
106.
▲
by
jasfi
2y ago
So like accelerators before YC? No thanks.
107.
▲
by
jasfi
2y ago
I think the real pain point you're hitting on there is that people feel like they don't get selected when they deserve to be. While there are those who don't deserve it but get selected anyway.
108.
▲
by
jasfi
2y ago
The AI coding tools I see today help to make coders somewhat more efficient. That isn't something that's very visible in the software landscape as a whole.
109.
▲
by
jasfi
2y ago
Circling around the idea in a response describes what I see in a lot of LLM output quite well. I haven't tried o1 myself, but it does seem to fix that problem.
110.
▲
by
jasfi
2y ago
More experienced software engineers can usually make the right design decisions that don't mean rebuilding when/if it comes time to scale things. This is usually about avoiding bad decisions because they know better.
111.
▲
by
jasfi
2y ago
I'm aiming for writing full-scale web apps with LLMs, mostly autonomously. The advice I got was to talk to customers, because this could be something nobody wants.
112.
▲
by
jasfi
2y ago
They fail operationally at serving the customer in those cases. But if the % of that happens is low enough, then strategically it doesn't affect them.
113.
▲
by
jasfi
2y ago
These enhance effects typically use guesswork. That would be useful to gain hints, but not to know for sure what the original looked like.
114.
▲
by
jasfi
2y ago
When companies can't tell you what they believe you did wrong, and point to a vague list of things it could have been, they fail.
115.
▲
by
jasfi
2y ago
I know someone who was greatly helped by acetyl-L-carnitine in that regard. Also, do everything you can to be healthy.
116.
▲
by
jasfi
2y ago
I was also a bit taken aback by the 2 space indentation. But after a bit of practice I got used to it.
117.
▲
by
jasfi
2y ago
Nim is a great systems language, and should be more popular.
118.
▲
by
jasfi
2y ago
I use the Options module which has a none/some check. None is the absence of a value. You can test for this quite easily and I see it as a feature, not a bug.
119.
▲
by
jasfi
2y ago
https://www.youtube.com/@CMUDatabaseGroup They publish their latest course videos every year, during the year. Andy Pavlo is highly knowledgeable about the field.
120.
▲
by
jasfi
2y ago
I agree. The reasoning is there, and becoming more capable every year (across the various models). It's easy to look for limitations, but what was once glaring problems are now much more subtle.
More ›