jsnell

21 ms·

451.
by jsnell 1y ago
Their weekly active user number is 10% of the estimated total number of Internet users (and growing rapidly). It's quite hard for that to happen with a high churn rate, because they'd very quickly run out of new users. For the num
452.
by jsnell 1y ago
I make it 8.9% with a binomial test[0]. I rounded that to 10%, because any more precision than that was not justified. Specifically, the results from the blog post are impossible: with 200 samples, you can't possibly have the claimed 5
453.
by jsnell 1y ago
The mode wasn't added after the announcement, Aider has had it for almost a year: https://aider.chat/HISTORY.html#aider-v0320 This benchmark has an authoritative source of results (the leaderboard), so it seems obvious
454.
by jsnell 1y ago
The 73% on the current leaderboard is using "diff", not "whole". (Well, diff-fenced, but the difference is just the location of the filename.)
455.
by jsnell 1y ago
That's not a lot of samples for such a small effect, I don't think it's statistically significant (p-value of around 10%).
456.
by jsnell 1y ago
https://aider.chat/docs/leaderboards/ shows 73% rather than 69% for Gemini 2.5 Pro? Looks like they also added the cost of the benchmark run to the leaderboard, which is quite cool. Cost per output token is no lon
457.
by jsnell 1y ago
No, it was funded by Google. Advertisers paid money for Google for totally unrelated services. Google invested that money in a number of ways. One of them was to build this very profitable non-advertising business. The advertisers didn&#x27
458.
by jsnell 2y ago
Because their main selling point is to run the copies concurrently with the original.
459.
by jsnell 2y ago
Obviously not, since you could not deliver it. It seems that you maybe don't realize what CFS is in this context, and are thinking of something else that you could just "set up"? What jeffbee is talking about is Google's
460.
by jsnell 2y ago
> Each step (though a little confusing to parse) does make sense in context. The proposed intermediate answers don't make sense, and they're the actual puzzle. You must answer all the intermediates correctly, you can't ski
461.
by jsnell 2y ago
There are almost certainly ways to fine-tune the model in ways that make it perform better on the Arena, but perform worse in other benchmarks or in practice. Usually that's not a good trade-off. What's being suggested here is tha
462.
by jsnell 2y ago
It's not actually solving the puzzle, for one. Like half the steps are obviously nonsense.
463.
by jsnell 2y ago
That's amazing, congrats!
464.
by jsnell 2y ago
> When he drops massive tariffs on everyone at once, it’s intended to force people to the table to negotiate with an uncomfortable starting position for them. It might be the intent, though there is no proof of it, just like there's
465.
by jsnell 2y ago
> And while risky, currency devaluation may follow later to make imports cheaper without lifting tariffs. Devaluation means a decrease in the value of the currency, making imports more expensive, not cheaper. That your source can make su
466.
by jsnell 2y ago
All of your pro-tariff points are ludicrous in the context of what actually happened, and do not reflect reality. This is not some negotiation tactic. We know it's not a negotiation tactic, because: there are no stated goals for negoti
467.
by jsnell 2y ago
Needing less memory for inference is the entire point of quantization. Saving the disk space or having a smaller download could not justify any level of quality degradation.
468.
by jsnell 2y ago
(2023) Original discussion: https://news.ycombinator.com/item?id=34657478
469.
by jsnell 2y ago
I'm not convinced. Password spraying attacks have a very low yield, and need high volume to compensate. 24 attempts per week is just not viable.
470.
by jsnell 2y ago
What you're describing is typically called credential stuffing. Password spraying is a distinct form of brute-forcing, where there is no link between the password and the user, and the yield is coming from the password being common.
471.
by jsnell 2y ago
Yes, but it would mitigate the issue almost completely since the app (and thus the address) would not be discoverable by new users.
472.
by jsnell 2y ago
That's a bit misleading. You can unpublish the app[0], which hides it in the Play Store and prevents new installs. It just doesn't remove the app from the people who already have the app installed, and I think they'll be able
473.
by jsnell 2y ago
No. If that interstitial is working, it's only working due to obscurity, and the moment this system becomes even slightly popular it'll become worthless. Proof of work is not a viable defense -- it's basically impossible to t
474.
by jsnell 2y ago
I guess I don't see which part you're finding objectionable. What you wrote initially was verifiably incorrect. It's not clear whether you made an honest mistake and just can't admit it now for some reason; whether you i
475.
by jsnell 2y ago
Calling this "Steam on Xbox" is A-grade clickbait.
476.
by jsnell 2y ago
You've moved the goalposts from "logging" to "inspecting". But pretty obviously checking for password reuse does not require logging, or otherwise storing, the passwords. Again: your claim was that Cloudflare had ad
477.
by jsnell 2y ago
The claim you made was "they just admitted to logging the usernames and passwords". The fact is that they did not admit to that. If you think otherwise, please quote the actual admission. I don't know how you think "this
478.
by jsnell 2y ago
Your source does not support the claim of the EU having announced a fine. It's an anonymous source saying that the EU will fine Apple and Meta later this month. You'd expect any fines to be announced along with the corrective meas
479.
by jsnell 2y ago
Well played by Apple, I guess. I really thought there would have been consequences for so blatantly ignoring the DMA. Instead they got to do so for over a year for absolutely free, and since I'm seeing no kind of timeline for implement
480.
by jsnell 2y ago
The link does not provide any supporting evidence for your claim about them logging passwords. Nor does the Cloudflare blog post that's the source. It seems pretty good that we aren't discussing something that you've just mad

More ›