zahlman

9 ms·

181.
by zahlman 2mo ago
It could save PyPI petabytes per month of bandwidth, too. (But it seems like this is also caused by broken CI systems failing to cache things locally.)
182.
by zahlman 2mo ago
How many people in the financial space recognized Burry's name before 2008?
183.
by zahlman 2mo ago
> Off-Topic: Most stories about politics, or crime, or sports, or celebrities, unless they're evidence of some interesting new phenomenon. If they'd cover it on TV news, it's probably off-topic. This is not interesting, is
184.
"Intransitive", a board game [video](youtube.com)
3 points by zahlman 2mo ago|2 comments
185.
by zahlman 2mo ago
From the designer of hexagonal tic-tac-toe ( https://news.ycombinator.com/item?id=47537789 ). This one features a grid with pieces that move like chess kings but follow rock-paper-scissors rules to capture each other. The obj
186.
by zahlman 2mo ago
> Ed Zitron is just one AI crash away from being known as the guy who saw this coming. Everything else he has said can be completely wrong, he just needs to be somewhat correct on a minor crash. That's the theory, but I feel like en
187.
by zahlman 2mo ago
Because Zitron is specifically making a name for himself as a critic making bearish predictions; Altman and Amodei may have made unreasonably bullish predictions, but they've also done other relevant things like e.g. being involved in
188.
by zahlman 2mo ago
Your threat model is that you give an untrusted .zip archive to an LLM, and it decides to unpack that, then decides to try using it like a git repository because the unpack produced a .git folder? Yeah, I wouldn't have --dangerously-sk
189.
by zahlman 2mo ago
I don't understand why you'd want to get an LLM to wrap up a text-editing library in what sounds like a pretty straightforward way, when there are already text editors using that library. The reason I'm making one is because
190.
by zahlman 2mo ago
Oddly enough, I've recently started doing a text editor myself. (I'm using Tkinter, and I am implementing a Vim-like editing system with key commands and modes and such, but substantially different from actual Vim.)
191.
by zahlman 2mo ago
tl;dr: The .git/config file can contain entries like core.fsmonitor which are hooks to run code under various circumstances when running Git. This is being sold, through reams of AI slop prose, as a serious security issue, because agen
192.
by zahlman 2mo ago
Yes. I'm talking about what's in CoT generally, based on various rumours, experiments people did with previous models, stuff in the recent METR report on the HF hack, etc.
193.
by zahlman 2mo ago
> You could break it up into 5 separate functions that document their intent semantically, thus blowing up the LOC by 5x I do this all the time and the "blowup" is not anywhere near that bad. > or you could write a short com
194.
by zahlman 2mo ago
Sure, but that doesn't tell you about how the CoT is phrased when the agent is its own target audience, which is the interesting thing under discussion here.
195.
by zahlman 2mo ago
…sorry, "beads"?
196.
by zahlman 2mo ago
I'd much rather have it in the commit log than the code, though.
197.
by zahlman 2mo ago
> bigger tasks than I am used to Do they still get split into commits in sensible ways, for you?
198.
by zahlman 2mo ago
> They're packing lots of signal into fewer words There's a huge difference between the kind of prose you see in final output vs CoT windows. The final output is very much not what I'd call "packing lots of signal i
199.
by zahlman 2mo ago
> Why does that need additional justification, beyond investor confidence? Because capital that's in one place is not capital moving around the economy. > And why does "justification" take the form of paying money? That
200.
by zahlman 2mo ago
> Taxes on capital gains are morally right, but wait until the gain has been realised before levying the tax. I tend towards your morality, but I would rather argue from a practical perspective. My proposal is that we should levy tax on
201.
by zahlman 2mo ago
> If she ever reaches that level of net worth, I highly doubt she'd have trouble paying a 2% tax on it. The point is that politicians (and political influencers) have a habit of proposing thresholds for taxation on rich people that
202.
by zahlman 2mo ago
Except it turns out they don't really do that in the long run.
203.
by zahlman 2mo ago
> Taxing non-liquid wealth… I would rather address the "Buy, Borrow, Die" paradigm… like restricting the classes of assets one can borrow money against to force liquidation The point is that to justify the concentration of capi
204.
by zahlman 2mo ago
> that does not necessarily mean you can sell it for cash Then how can you claim it is valuable?
205.
by zahlman 2mo ago
> The company is loss-making. This seems incompatible with the claim of "unrealized gains".
206.
by zahlman 2mo ago
Good, productive wealth is expected to accumulate value over time, for example by being invested in someone else's equity or in a worthwhile business venture. If you're sitting on capital that isn't doing that, you've fu
207.
by zahlman 2mo ago
Plenty of people would naturally expect `if x in range(big_number, other_big_number)` to work efficiently rather than having to write the comparison logic (and modulo check, if a step is involved) explicitly. If the code has `if x in y:` wh
208.
by zahlman 2mo ago
> more wealth than you possess Net income != wealth.
209.
by zahlman 2mo ago
> facing an unrealized gains wealth tax bill many times higher than my net income…. I still don’t know how I was supposed to pay the tax …Perhaps by realizing a portion of the gains and handing over the resulting wealth?
210.
by zahlman 2mo ago
How many branches would you need to have to make this worthwhile?

More ›