vlovich123

6 ms·

1.
by vlovich123 3d ago
For what it’s worth the GCC backend for rustc is making progress and rustc more generally supports a lot of embedded stuff already
2.
by vlovich123 3d ago
One main difference is that pyo3 is safer and easier to use to build said modules and have them work correctly without issues. Thus modules which hadn’t been worth it before to build in this manner suddenly become so
3.
by vlovich123 4d ago
The 80186 didn’t really introduce anything new architecturally. It’s basically an 8086 with a few more chips bundled on-die. 80286 however introduced protected mode, expanded the address size to 24 bit, hardware enforce memory protection, m
4.
by vlovich123 4d ago
Three - after 8086 you had 80286, 80386, 80486 and then Pentium (superscalar).
5.
by vlovich123 4d ago
Your own link contradicts itself. If Google funding was so critical to the plant staying open at all, how come they were already planning a 38MW expansion independent of that? That doesn’t sound like it’s on the brink of closure. You should
6.
by vlovich123 4d ago
Which is irrelevant to the article which talks about the amount of energy used.
7.
by vlovich123 4d ago
While mostly true, don’t discount people looking for work that offers them familiar tools. The cheap public pans are a way to train a labor force + create upward demand from people trained on those tools to influence big corporate purchase
8.
by vlovich123 4d ago
American cities also generally seem to have good 5G coverage with 4G frequencies being repurposed for 5G and 4G overall being put on life support as it’s phased out
9.
by vlovich123 4d ago
Installing renewable energy doesn’t change the amount of energy required. The shift to Asia is almost certainly the bulk of it
10.
by vlovich123 4d ago
> I don’t think a manufacturer should be able to lock you out of reading diagnostics on your own devices, or that anyone should be prevented from reverse engineering what those values mean. But they do do that. Also what’s the difference
11.
by vlovich123 5d ago
Sure I was assuming it wasn't a 1 year outlay. Still, even over several years that's a lot for just software - you're talking about $250M/year in development costs split over 4 years across 600 developers. I'm assum
12.
by vlovich123 5d ago
What do you think batteries are if not baseload generation? Batteries + renewables compete with nuclear for baseload. They take up substantially more land mass for the same amount of power though.
13.
by vlovich123 5d ago
> Those decisions would otherwise have been made year by year, but the long-term deal with Google ensures Loviisa will keep running through the end of its operating license It's important to be able to read through PR in planted pie
14.
by vlovich123 6d ago
Source? The provider had requested to extend the lifetime of that particular reactor to 2050 2 years ago (something they’d started considering 8 years ago). Now maybe they did this contingent on expecting this Google deal to materialize, bu
15.
by vlovich123 6d ago
Nuclear power plants have a huge lead time and require government approval, something the Finnish government declined to approve (Fortum had requested to build a third reactor). Solar and wind are fine but they’re not usable as a replacemen
16.
by vlovich123 6d ago
If I consume all the clean energy your country can produce with my data center, explain how cap and trade blocks a country from using other sources - I don’t think anyone will like news stories of people freezing to death in the winter.
17.
by vlovich123 6d ago
Where did you read this? What I read is that Fortum was considering extending to 2050 as early as 2018 and finalized those plans 2 years ago. I read nothing that indicates the survival of this plant is dependent on this Google deal. https:
18.
by vlovich123 6d ago
It’s unclear. It helps Google with their marketing that they’re a net zero carbon company. But if what’s happening is they’re displacing previous nuclear power users onto dirtier, that’s not really reducing carbon and it’s making it seem li
19.
by vlovich123 6d ago
You’re assuming a large enough supply that can absorb demand. This is true in China but not generally in the West. If supply was just in time constructed so you maybe have only like 10-30% extra vs peak, the 50% of a nuclear power plant shi
20.
by vlovich123 6d ago
Based on Grok identifying the company correctly, that would amount to ~3% of annual revenue and 30% of profits which seems insane to me for the cost of a single project. Anyway, even if that were true, rewrites are becoming drastically chea
21.
by vlovich123 6d ago
Yeah and Astra is much better still
22.
by vlovich123 6d ago
You benchmarked a really old heuristic hash in Murmur and xxh64. xxh3 is ~1.6x to 1.8x faster than xxh64 - extrapolating that to 5.7 in your paper puts it at ~18-27% faster. WyHash and UMash sometimes report faster numbers. Even xxh3 isn&#x
23.
by vlovich123 6d ago
Have you tried the latest motion cues in iOS/Android? They’re supposed to combat this.
24.
by vlovich123 7d ago
Would this be applicable to fast hashes like WyHash and xxh3 or are those not using polynomials? Is this mainly for faster cryptographic hashes?
25.
by vlovich123 7d ago
The first new resolution for an iPhone was 2012 which also saw the introduction of Auto Layout. The second new resolution was in 2014 which added size classes. It's entirely possible that the 2013 game was built before they actually st
26.
by vlovich123 8d ago
Generally infill is significantly faster than inference due to batching. Is that not the case here for some reason?
27.
by vlovich123 8d ago
I think you may be underestimating how difficult a text search over their data is. They may have to build new mechanisms to do this. And what you really want is also an attribution of how much of a contribution a given corpus made which is
28.
by vlovich123 8d ago
Not sure it’s that strong. The prompt gives the presumption that this matters. Not necessarily a training issue vs the prompts being poorly written and the results being inherent in the bias they carry
29.
by vlovich123 8d ago
Unless they know exactly the researcher’s account, they may not know in their end if he had the setting to let them train on his chat logs. They also probably don’t know if he had any correspondence on any forum where he may have discussed
30.
by vlovich123 8d ago
Of course you do if you're a) only given a partial solution b) racing against someone else using a competing AI. The open question was whether their LLM got the nudge in the right direction because it got access to the chat somehow (e.

More ›