anuramat

7 ms·

151.
by anuramat 11mo ago
I really like how Kagi does it - summary only appears if you finish your query with a question mark, or if you press "q" after loading the results
152.
by anuramat 11mo ago
> Are they "decisively" choosing it I'm pretty sure chatgpt and perplexity are "opt-in" in this sense > nothing to do with language models how do you think translation works?
153.
by anuramat 11mo ago
reminded me of "tensor is a bunch of numbers that transform in a certain way"; this should be illegal to teach, especially in physics
154.
by anuramat 11mo ago
> not some true disorder there's a tool on your desk that might help you solve your problem; what does it matter if the problem is an "appropriate response of your body"? so is pain/anxiety/diarrhea
155.
by anuramat 11mo ago
> so much slower unless you're working on 10 features in parallel with smaller models in best-of-4 groups
156.
by anuramat 11mo ago
1. my point is that "thinking" is easier to say than "composition of parameterized nonlinear functions trained by stochastic gradient descent with reinforcement learning on top". misnomer or not, it's not even ambig
157.
by anuramat 11mo ago
I thought you still need to implement the patching on your side? judging by < https://platform.openai.com/docs/guides/tools-apply-patch >
158.
by anuramat 11mo ago
so why not give mozilla the benefit of the doubt? why assume that it's necessarily gonna be a slopfest? I'm not exactly optimistic, but it could be good, I do want those features, and I'm not gonna switch to something like
159.
by anuramat 11mo ago
> good idea for them to make the features opt in true, but that's not the same as "nobody wants AI"; them fucking it up once shouldn't mean that they should drop the entire idea; by this logic they should stop making
160.
MiniMax M2 Coding Plan(platform.minimaxi.com)
1 points by anuramat 11mo ago|0 comments
161.
by anuramat 11mo ago
counterpoint: I want AI in firefox there's so much stuff that could get much better if they invested more in AI features -- tab grouping, translation, ad blockers; why are people so triggered? because it might end up being bad?
162.
by anuramat 11mo ago
by this logic sockets are also non-unix
163.
by anuramat 11mo ago
what's the point of apply_patch and shell tools though? can't you just define your custom tools with exactly the same behaviour, since you're implementing the actual execution on your side anyway? sounds like vendor lock in f
164.
by anuramat 11mo ago
sounds like this is just a new snapshot, so I don't think anything changed (upd: anything about their marketing I mean)
165.
by anuramat 1y ago
two questions: 1. do you ever point out that you can't actually mine bitcoin with a pickaxe? 2. what made you think that the parent comment somehow implied that it "actually thinks"?
166.
by anuramat 1y ago
stylistic preferences are pretty much the ONLY thing you could discuss (in the context of LLMs) that actually has anything to do with (natural) language in the first place; how is having preferences an "improper use of langauge"?
167.
by anuramat 1y ago
do you ever get tired of pointing out that a large language model is a language model? UPD I do that as well when explaining to my relatives why I don't care what ChatGPT thinks about $X, but also they're not on HN
168.
by anuramat 1y ago
> codebase they are started in what about the dependencies? you could just clone them as well (which is what I do occasionally), but deepwiki is faster (for indexed repos) and free
169.
by anuramat 1y ago
how is codex on this list though? "agent" is also alive for now, and I doubt it's gonna go down any time soon
170.
by anuramat 1y ago
> inject a prompt based on the description how are skills different from SlashCommand tool in claude-code then?
171.
by anuramat 1y ago
for me its the speed; eg cerebras qwen coder gets you a completely different workflow as its practically instant (3k tps) -- feels less like an agent and more like a natural language shell, very helpful for iterating on a plan that you them
172.
by anuramat 1y ago
why reiterate the invariant?
173.
by anuramat 1y ago
even if it works as described, I'm assuming it's extremely model dependent (eg book prerequisites), so you'd have to re-run this for every model you use, this is basically poor man's finetuning; maybe explicit support fr
174.
by anuramat 1y ago
is better slop a bad thing somehow?
175.
by anuramat 1y ago
take a look at the replies to https://news.ycombinator.com/item?id=45112846 eg https://www.youtube.com/watch?v=sQYXZCUvpIc and https://www.youtube.com/watch?v=EL7Au1tzNxE PS, what was your
176.
by anuramat 1y ago
Of course, but why post it publicly ? I personally keep all my accounts private even though I am not a 13 years old girl, so I don't get why you would expose your children like that With or without ads, there's gonna be a certain
177.
by anuramat 1y ago
> 13-year-old ... schoolgirls in short skirts with either bare legs or stockings why would one post these in the first place?
178.
by anuramat 1y ago
the address is fake though
179.
by anuramat 1y ago
slightly offtopic, but the challenge might be leaking emails: just got an email from `alfredw@listenlabs.fyi` (note the TLD): > I'd like to connect you with our team to hear about your solution. > 1) can you let me know availabil
180.
by anuramat 1y ago
I guess "claude code but 20x cheaper" sells better; you can still use their subscription with other agents, just use their anthropic-compatible API

More ›