jsnell

5 ms·

31.
by jsnell 2mo ago
Good grief that thread is sad, with it ending by Gary Marcus asking for the crank to be taken seriously.
32.
by jsnell 2mo ago
There's no business in straight-up news aggregation. Google News is purely a public service (it isn't monetized) and a liability (see the regular attempts by governments to make Google pay for news). Anybody with any kind of busin
33.
by jsnell 2mo ago
Now, now, let's be fair. I don't think LLMs are used for writing or editing the articles in any way. (Each reader can decide themselves whether what you wrote is unfair to Zitron or to LLMs.)
34.
by jsnell 2mo ago
The original was an actual quote? But "had" still doesn't mean what you are implying: once the model solved the problems and the solutions were verified, the problems weren't open any more, so a later description using t
35.
by jsnell 2mo ago
> analysts have estimated that it will take $2 trillion a year in revenue to pay for the infrastructure that has already been built I doubt any credible analyst has claimed that. What's the total AI capex that's already been sp
36.
by jsnell 2mo ago
> A back-of-the-envelope calculation finds that covering AI capex through identifiable AI income requires revenue on the order of $2.5trn per year, Ok, so what is the botec? They don't say, but it's probably something like &quo
37.
by jsnell 2mo ago
Isn't it up to the open model advocates and publishers to come up with the solutions for making them safe? Like, there's three plausible arguments about safety of open models: 1. Any concerns are fake news. Open models will always
38.
by jsnell 2mo ago
The publisher, in this case Postbank, decides what countries to allow installs from for an app. Google just enforces the policy the publisher sets.
39.
by jsnell 2mo ago
It's very hard to make sure none of the outputs are ever made available to third parties. Source code can end up widely distributed (e.g. client-side js, open source). Prose will frequently get shared across organization boundaries (e.
40.
by jsnell 2mo ago
Some of his numbers are rebutted all the time. It doesn't matter, because he will just ignore the rebuttals and just keep repeating the invalid or irrelevant numbers. For example, a few months ago he did some bad math with a broken spr
41.
by jsnell 2mo ago
You're biasing for quantity of argument vs. quality of argument. That's Zitron's MO: writing 10k word rambles that cover 20 topics, ostensibly with tons of links to sourcing, but with most of the sources not actually saying w
42.
by jsnell 2mo ago
Early in terms of there being no sign of actual signs of decay of their business. It's less than a week since Google reported literally the highest quarterly profits by any public company ever[0]. This is while growing at 25% YoY, whi
43.
by jsnell 2mo ago
Relevant article + discussion from earlier today: https://news.ycombinator.com/item?id=49014034
44.
by jsnell 2mo ago
The exact claim I was replying to "It does that by increasing the cost of scraping, not by preventing it entirely." In that context, quantifying the cost isn't theoretical nitpicking! Anubis works to the extent it does becaus
45.
by jsnell 2mo ago
Except it does not actually increase the cost of scraping meaningfully. Compute is really cheap. The compute for minting an Anubis cookie will cost less than a thousandth of a cent even assuming the attacker uses the same JS implementatio
46.
by jsnell 2mo ago
Unbearable AI slop, flagged.
47.
by jsnell 2mo ago
Your title leaving out the "innocent" part makes it a stunningly inaccurate summary.
48.
by jsnell 2mo ago
Uhh... This is not a popularity chart. It is an aggregate of benchmarks. It baffles me that somebody would write something that aggressive from that deep a level of confusion.
49.
Scanning for Pangram Errors(veryfineprint.substack.com)
73 points by jsnell 3mo ago|43 comments
50.
by jsnell 3mo ago
[flagged] means flagged by users; if it was done by the system, it'd be [dead]. I can't speak for others, but the reason I flagged it is that the number is untrustworthy and absurd. This is not an isolated case, Statcounter has th
51.
by jsnell 3mo ago
Your theory is pretty silly. If the outcome is doom, nobody cares whether you were right, because humans are dead or irrelevant. The asymmetry in the reward function isn't there. In general the rewards for public doomerism seem low, ex
52.
by jsnell 3mo ago
It doesn't make the economics any different. In a browser environment, you're maybe looking at the acceptable lease being 100MB for 1 second. Much more than that, and you start hitting limits of what browsers will let you do on lo
53.
by jsnell 3mo ago
Proof of work does not scale. It trades something fungible and incredibly cheap (CPU) for something incredibly expensive (user-visible latency). There is no set of parameters where the cost is going to be a meaningful deterrent to any kind
54.
by jsnell 3mo ago
My point is that Kelley did not argue that what Bun does isn't really fuzzing. He wrote that the post's claim is a fabrication. But that claim is really specific, and to evaluate whether it is true it doesn't matter what Kell
55.
by jsnell 3mo ago
I don't understand what distinction you're trying to draw here. The very specific claim[0] in the Bun blog post that Kelley is calling a fabrication was: > We fuzz Bun's runtime APIs 24/7 using Fuzzilli, the JavaScrip
56.
by jsnell 3mo ago
> that this blog was so clearly written by LLM's is offputting for some reason It doesn't read at all AI-generated to me. What section do you think is? (Pangram is very good at distinguishing between AI-generated and human text
57.
by jsnell 3mo ago
AI-generated slop, flagging.
58.
by jsnell 3mo ago
This is AI-generated slop, flagging.
59.
by jsnell 3mo ago
Just to nitpick, requiring a free account is substantially different from a paywall.
60.
by jsnell 3mo ago
I'd say "entirely", not "partially". Flagged the article.

More ›