ben_w

31 ms·

571.
by ben_w 3mo ago
> Yeah. 5 ly sounds much safer. Well yeah. I clearly could've phrased this better though: I was aiming for "here is an example distance which *is* bad", not as it appears to have been understood "here is the distance
572.
by ben_w 3mo ago
We know what something with about 2 solar masses looks like at about 5 light years, because there is such a thing a bit closer: the Alpha Centauri system. So (and again, this is very rough approximation) I mean AU: consider the high speed o
573.
by ben_w 3mo ago
The bad news is that if a black hole is not primordial, the minimum mass is high enough (2-4 M_sol) that it doesn't so much "pass through our system" as "we pass through its". I'm not 100% confident, but rough
574.
by ben_w 3mo ago
> I suspect the horses are healthier and stronger than in the 1700s too. I'm not sure I understand your point? Horses? I doubt it. They're still infamously fragile creatures today, even with penicillin and vaccinations existing
575.
by ben_w 3mo ago
That, and even Apple's been copying others for a while now.
576.
by ben_w 3mo ago
> In this repository, multiple files are close to 10K LOC, one file contains a switch statement that has so many case statements it spans more than 1000 lines, and lots of other fun stuff. Pft. Over multiple cases? I've seen 1000 li
577.
by ben_w 3mo ago
On a scale of A+ to F, Anthropic are a C+ and OpenAI a C. Such mediocrity makes them the best of the mostly-terrible bunch. https://futureoflife.org/ai-safety-index-summer-2026/#scorec...
578.
by ben_w 3mo ago
> The fact that Facebook has so much experience taking advantage of people's private data is one of the reasons I believe them when they say they won't be doing it when you pay them for that service. To me, their history sugges
579.
by ben_w 3mo ago
Email addresses are like other contact information: people don't like having to update it (on either side of the link), it is sticky and people will keep sending you messages on the old address long after you stop sending from it, so i
580.
by ben_w 3mo ago
How is me installing that going to help prevent Facebook hosting unlawful adverts? I barely visit the FB feed these days, about half of those visits being a notification that turns out to be someone's birthday. Unlike FB themselves for
581.
by ben_w 3mo ago
Interesting. That won't work for me directly (iOS), but it made me search around and find this, which has xcancel in one of the screenshots: https://apps.apple.com/de/app/redirectweb-smart-url-router/i...
582.
by ben_w 3mo ago
Aye. My (literally, and self-proclaimed) communist former partner noted left-wing infighting. The Monty Python Life of Brian scene about The Popular Front of Judea ("he's over there") are, according to her, totally accurate i
583.
by ben_w 3mo ago
if someone shares me a link to x.com, it is literally faster for me to edit the URL to xcancel, even on my phone which makes this awkward, and load that, than to wait for the original URL to load. xcancel does not have a login barrier, nor
584.
by ben_w 3mo ago
> I'm wondering what people expect these companies to do? Human moderators? Sure, but Meta products have approximately half the worlds population using them. At this scale it's pretty mind bogglingly difficult to moderate down
585.
by ben_w 3mo ago
There are basically always filters of some kind, though not necessarily the ones we want. The bigger the target that filter represents, the more effort is spent breaking it.
586.
by ben_w 3mo ago
> Cloud, Gmail, YouTube, Android, Chrome Cloud and gmail, I agree. I don't think YouTube can survive if GenAI keeps going like this. Android, same, but on a different timescale and for different reasons. The Play Store (and all othe
587.
by ben_w 3mo ago
> It remains to be seen if public investors are as ambitious as the engineers. And what politicians do when "Damn immigrants coming here taking our jobs" turns into "Damn remote workers, staying right where we i
588.
by ben_w 3mo ago
> why neutron moderation is needed What you learned, was it more like: 1. "You need to slow down neutrons so they can react" or 2. "Here's the graphs of how the neutron absorption and scattering cross sections vary wi
589.
by ben_w 3mo ago
By that standard, I could've been an astronaut at that age. Newton's laws of motions are not hard. Making a rocket that doesn't kill the occupant, is.
590.
by ben_w 3mo ago
> Starship has as much chance of working to spec as the Theranos machine. Don't get me wrong, I'm skeptical about SpaceX's business model (I really must make more time to work on that blog post about orbital compute…), but
591.
by ben_w 3mo ago
Generally. Sometimes it still did, in my experience.
592.
by ben_w 3mo ago
> services that they operate that have high demand. "high" doing heavy lifting there, I think. SpaceX has a lot of demand for rocketry… from itself, in the form of Starlink. Starlink has segments where it's a near monopoly
593.
by ben_w 3mo ago
> would you have to give access to modern hardware in the experiment for it to use? At a minimum, yes. IIRC, the sum total of all compute manufactured over history only reached the minimum needed to train an OK LMM in the mid 00s. > H
594.
by ben_w 3mo ago
If you gave GPT-2 a question and ended with a "?", it might answer, but also it might write several more questions in a similar category. IMO, the mechanism isn't the important thing, the behaviour is. If you look at the step
595.
by ben_w 3mo ago
> One of the experiments I've heard proposed around here is to somehow create an LLM from all text up to 1980 or 1990 and see if it can get back to making itself. Could be, but preventing leakage from more modern stuff can be challe
596.
by ben_w 3mo ago
Lots of things changed, GPT-2 is small (1.5e9) and is also a base model, so it is only doing next-token/autocomplete rather than prompt-response like even the first ChatGPT-3.5 was doing.
597.
by ben_w 3mo ago
Between the financials on one side, vs. graveyard, innovation by buying startups, dropping "don't be evil", and ads everywhere on the other, I don't know if you're suggesting this is good or bad.
598.
by ben_w 3mo ago
Why would open models change this argument? Which model is "best" changes too fast to be worth caring which particular model is the best, even on the timescale of making youtube videos. If anything, my initial expectation would be
599.
by ben_w 3mo ago
My memory of 24 years ago is a bit fuzzy, but I don't recall any ads on XP; the main complaint at the time was the UI was "fisher price".
600.
by ben_w 3mo ago
https://www.hitzewarnungen.de/index.jsp https://weatherkit.apple.com/alertDetails/index.html?ids=081... I'd describe it a little differently myself.

More ›