qsort

11 ms·

211.
by qsort 1y ago
The problem with this guy is that it's always the same 40 minutes on loop.
212.
by qsort 1y ago
> I kind of hope you're right. I couldn't care less if big tech gets knocked down a peg, but in many quarters the AI boom is what's keeping the lights on. A market correction of that magnitude would mean a lot of pain for
213.
by qsort 1y ago
Yeah, and? There are two separate conversations, one about capabilities and one about what happens assuming a certain capability threshold is met. They are p(A) and p(B|A). I myself don't fully buy the idea that you can just naively ex
214.
by qsort 1y ago
Yes, precisely. OP, you're overthinking this. As a consultant I talk with a lot of people, nobody whose opinion you care about will think less of you for saying exactly that. An intermediate answer you can give is something like "
215.
by qsort 1y ago
I don't think we are in a phase where we can confidently state that there's a correct answer on how to do development, productivity self reports are notoriously unreliable. At least personally, the reason why I prefer CLI tools li
216.
by qsort 1y ago
I could be wrong about this, but it feels like Cursor is less and less compelling with better models and better CLI tools popping up. Are the plan limits generous enough that it's worth a spin? Again, I haven't used Cursor in a wh
217.
by qsort 1y ago
This works very well for contractors, less so for full-time employees. You can't just fire somebody on a whim, at least not for free, not even in the US, let alone in most of Europe. To be clear, I'm not saying worker protections
218.
by qsort 1y ago
I'm surprised nobody is working on a "git for normal people" kind of software. Part of the reason why Codex or Claude can work the way they do is that even if they screw up you always have a clear picture of what were the cha
219.
by qsort 1y ago
It's free, it has the same features as commercial servers, it's open source, developer-friendly, with no ads (not even on free accounts) and a transparent corporate structure under French law. It's almost ridiculously good. D
220.
by qsort 1y ago
"Proceed with caution" seems to be the overwhelming consensus, at least with models having this level of capability. I commend the author for having the humility to recognize the limit of their capability, something we developers
221.
by qsort 1y ago
I wouldn't read too much into this particular launch. There's very good stuff and there are the most inane consumery "who even asked" things like these.
222.
by qsort 1y ago
I wonder if first-party offerings like Codex and Claude will follow suit. Most "agents" are utter nonsense, but they cooked with the CLI tools. It'd be a shame to let go of them.
223.
by qsort 1y ago
Ok, real talk. The implementation is very basic but it's fine, there isn't anything that's very wrong about it at a cursory glance. The choice of having "method style" calls as function pointers is extremely bafflin
224.
by qsort 1y ago
Eh, kind of. In a way, AI does not change at all the problem of having taste. There are more books you'll ever read, movies you'll ever watch, games you'll ever play, software you'll ever use. I remain completely unconvi
225.
by qsort 1y ago
How large? In many cases dumping to file and bulk loading is good enough. SQL Server in particular has openrowsets that support bulk operations, which is especially handy if you're transferring data over the network.
226.
by qsort 1y ago
I do expect this package to make connecting easier, but it was okay even before. ODBC connectivity via pyodbc has always worked quite well and it wasn't really any different when compared to any other ODBC source. I'm more on the
227.
by qsort 1y ago
At least from my point of view (in the industry, not academia) this is actually the opposite. Math graduates tend to be smart and humble and I respect them a lot. Sometimes it almost feels like math and physics are the last "real"
228.
by qsort 1y ago
> how in anonymous forums there's a lot more people pointing out that they think this is hype whereas when we wear our professional hats many of us join in Different speeches for different audiences. On HN, for all its faults, peopl
229.
by qsort 1y ago
I particularly like their usage of LLM-as-a-judge. They don't go "hey chatgpt, sort these from best to worst based on vibes", rather they extract a set of ground truths and check how the answer compares, a task that SOTA LLM
230.
by qsort 1y ago
If I had to guess, something related to floating point operations. FP additions and multiplications are neither commutative nor associative.
231.
by qsort 1y ago
Veritas used to represent abstract truth is not out of place. Obviously it assumes a different connotation in a Christian context ("Veritas vos liberabit" from the gospel of John being the obvious example), but it's not the o
232.
by qsort 1y ago
Was directed at TFA, not parent comment.
233.
by qsort 1y ago
Isn't that happening already? Half the usual CS curriculum is either math (analysis, linear algebra, numerical methods) or math in anything but name (computability theory, complexity theory). There's a lot of very legitimate criti
234.
by qsort 1y ago
I won't complain about a strict upgrade, but that's a pricy boi. Interesting to see differential pricing based on size of input, which is understandable given the O(n^2) nature of attention.
235.
by qsort 1y ago
Python still doesn't have tail recursion, and uses a small stack by default. I'll note that in modern imperative languages is harder than it looks to figure out if calls are really in tail position, things like exception handling,
236.
by qsort 1y ago
Regrettable, but did it take o3 mega pro to find out about real and nominal value? Even something a trivial as an iPhone is a far bigger purchase if you're not on a Bay Area salary.
237.
by qsort 1y ago
It's a phase. I used to try and customize everything, tiling window managers, custom color schemes, Arch, etc. Right now I'm on a Mac so vanilla I didn't even change the wallpaper.
238.
by qsort 1y ago
Thankfully that isn't a problem: we have scientific and reliable benchmarks to cut through the nonsense! Oh wait...
239.
by qsort 1y ago
The original comment was this: > So there's no ground truth; they're just benchmarking how impressive an LLM's code review sounds to a different LLM. Hard to tell what to make of that. The comment I replied to was: > Th
240.
by qsort 1y ago
No, they aren't. Most benchmarks use ground truth, not evaluation by another LLM. Using another LLM as verifier, aside from the obvious "quis custodiet custodes ipsos", opens an entire can of worms, such as the fact that ther

More ›