ben_w

16 ms·

301.
by ben_w 2mo ago
> but pretty sure even ChatGPT Free does that these days). It can do, but won't do so reliably; and the guessing it does if it's not grounding with a search is still like someone in a pub saying "I recon…". Admittedly
302.
by ben_w 2mo ago
Uranium-235 has a half-life of 704 million years. It emits so slowly that the main hazard is heavy metal poisoning, which you'd also get from the depleted uranium that's a waste product from enrichment. From the numbers I see on
303.
by ben_w 2mo ago
I suspect that would backfire immediately, as it would mean all stories are indirectly (via many people using this) filtered by what Gemini finds plausible. Purely as an example, when Trump was making the threats against Greenland, I tried
304.
by ben_w 2mo ago
> And that doesn't matter. If Meta is forced to issue 10% more shares to the plaintiff the value of all shareholders' shares fall 10%. And if the plaintiff dumps all the shares on a single day, they fall even harder. And why sh
305.
by ben_w 2mo ago
Given the C&D against the sane ways to read any X post without an account, what's this about?
306.
by ben_w 2mo ago
That's like a court ordering 10% of your home be put on the market in a way that never forces you to pay rent, when you're likely to live in it for the rest of your life. The wealth you're talking about is almost entirely bou
307.
by ben_w 2mo ago
Would if I could, but I think you're missing my point somewhat.
308.
by ben_w 2mo ago
The message being "screw the rules, I have money". The shareholders already can't remove Zuckerberg, he has 57% of voting control.
309.
by ben_w 2mo ago
What % of META (shares) are owned by Meta Platforms Inc.? Don't get me wrong, I'm happy for e.g. Zuckerberg personally to be ordered to hand over all his shares as any given penalty, but as stated what you describe sounds like a w
310.
by ben_w 2mo ago
The sibling comment, talking about Schrödinger, is correct for the wrong reason. The share price is essentially always fake, because the ticker price of any given symbol is in a superposition of all possible values until it gets traded. The
311.
by ben_w 2mo ago
> Why would you dumb down market cap to earnings? The point is "Market cap is BS". > Anyway, 6K for harming children at that scale remains ridiculous. One person managing to harm children at this scale would likely get a pen
312.
by ben_w 2mo ago
In the context of: Firefighters who deal with forest fires, police, governments, rescue services (here in Spain) are all posting their updates on X - because, like the users, they have no choice, because no alternative exists. It sim
313.
by ben_w 2mo ago
Market cap is generally the wrong measure. If you registered asf1254 Inc., and it was always just you, and it traded like Meta stocks trade, then "asf1254 Inc." would have a market cap ~6.2× whatever you're currently earning.
314.
by ben_w 2mo ago
You tell me. Most of these reports take that long to get published, or even longer. Sometimes I even see new-ish reports talking about 4o.
315.
by ben_w 2mo ago
> which is obviously a significant privacy complication. I'm sure the companies will default to the darkest of dark patterns, and the most privacy violating mechanism possible, but the law doesn't have to be written to require
316.
by ben_w 2mo ago
> Imagine being able to run GPT Sol at 1k tokens/second a year from now, at a 100x lower cost per token than now. Would that be useful? Given the current rate of change, it would be hard to guess either way. By some measures the cos
317.
by ben_w 2mo ago
Dark matter is mysterious enough to be compatible with both at the same time, I think. (Is a collisionless gas really even an "object"?)
318.
by ben_w 2mo ago
Normally people think of gravity as pulling on objects. You can instead think of it as pulling on the space those objects are in. A black hole happens when there is enough gravity that space gets pulled inwards somewhere, at at least the sp
319.
by ben_w 2mo ago
From the above stats, a 2.3% change to USA meat consumption, up or down, has the same impact as a 100% change to data centre water consumption, up or down. If I read the USDA stats correctly, US red meat and poultry production went up about
320.
by ben_w 2mo ago
While I agree fully with this part: > Datacenters should be required by law to pay to build adequate power generation and infrastructure to pay for their consumption, and ideally in a way where it isn't a ghastly eyesore. Fact is, a
321.
by ben_w 2mo ago
It'll work for now, at least. I fully expect that in my lifetime someone will offer a service of tissue-culturing just enough to make an inverse-cyborg* to handle all your, ahem, meat-space obligations. I mean, I'm already follo
322.
by ben_w 2mo ago
> Outside of the most niche domains like frontier AI development, are there software engineering questions that Sol 5.6/Fable 5 can't answer better than the median software engineer if explicitly asked? "Median" is ki
323.
by ben_w 2mo ago
I'll take your word for it, I think I'm in zero groups on Facebook, I only even look at the feed when I see a notification in Messenger; and I barely use Messenger, too.
324.
by ben_w 2mo ago
Coding < Development < Engineering. My claim is just that LLMs are basically fine now with the "coding" part. Not so much with development, definitely not with engineering.
325.
by ben_w 2mo ago
Out of curiosity, I just tried this: write a comment about how to use ChatGPT to write a game like you're a barely literate reddit user, short answer, maximum incoherency and the result was a pretty good simulation. I tried less
326.
by ben_w 2mo ago
I have also tried developing games with them. I find them very weak at these things. Often they can't figure out how to test the thing, so "does this attempted solution even work at all?" is its guess from reading the source
327.
by ben_w 2mo ago
Aye. This is true for many professions, I think. I'm trying to write a novel, have been for a while now; one thing I hear from professional writers is that one should only do this if you are prepared to re-read whole thing six times be
328.
by ben_w 2mo ago
> That sounds great until you realize that "hey Claude, read this Jira ticket and implement the feature in this code base" isn't really worth $200K/year. Nor even €50k/year. Two things are true: 1) The coding par
329.
by ben_w 2mo ago
> Why do models do this? > Likely the pretrain datasets are at fault They don't really know what hardware they're running on, or what load that hardware is under at the time you ask. I use these "human time required&quo
330.
by ben_w 2mo ago
While yes, their current reliability is too spiky to be relied upon for a lot of things: Any given failure is not inherent , they are all dependent failures; what is inherent (due to the SOTA in ML, perhaps or perhaps not the architectur

More ›