sigfig@sigfig

i saw a space ship blow up once and immediately decided i wanted to get in one

 sf
Joined November 2012
415 Photos and videos
Pinned Tweet
sigfig@sigfig
11 Nov 2024
gravity is a greater force than history
1
1
69
33,064
the honey is sweet. the pot is deep. the money is peat. and many will be burned
8
355
hahahahahaha how is the fuck is the erosion of institutionally-mediated consensus reality real just log off
1
1
17
525
opus 5.6 will complete the system of german idealism
Sebastian Thrun
@SebastianThrun
Sep 24
1/ Introducing PhilosophyBench from @StanfordAILab @StanfordHCI, the first independent, large-scale benchmark for evaluating AI’s philosophical capabilities. philosophybench.org
1
28
1,262
i think many people will be surprised to find this idea given any credence at all by an academic philosophy department, and it's important to note that it hasn't been. this is supported by oxford and stanford, which are not real
1
12
276
after checking i feel like i should hedge my stanford slander here, their letterhead is on the project but i do not believe any of the participants are faculty there
3
172
i would recommend teaching your children how to critically engage with derek parfit's "reasons and persons" before you allow them internet access. it is apparently a critically important life skill
22
537
it has been very interesting (and also very jarring) to see my own thoughts and writing become so much more structured, at the same time as the thoughts and writings of people i respect as academics and authors have become progressively more unstructured
2
2
35
1,026
i don't entirely know what to make of this yet, and i don't think it's any sort of moral emergency really, but i can see on your blog when you started using claude
1
2
13
267
the dm interface freezes regularly because its frontend state is stored in a wasm build of sqlite that locks for transactions every time a message is received in any chat
Paul Graham
@paulg
Sep 19
Remember how eagerly the press swallowed stories about how Twitter was going to crash after Elon bought it? How could they have believed that someone who could run rocket and car companies couldn't run a forum?
26
34
886
28,368
old twitter was a poorly managed project in its own way but they were in a completely different league than the current product. the current team regularly surprises me with their ability to invent entirely new ways to degrade user expectations
1
1
242
2,777
once every 50 years san francisco releases a potent psychotoxin
2
6
50
1,066
"no way to prevent flood" says only zone where this ever happens
1
22
877
i wish we could talk about the recent proof generation stuff without the millenarian nonsense. the existence of a "random counterexample generator" for arbitrary (or at least a large class of) theorems is legitimately very interesting in its own right
1
1
34
1,184
it's not a revolutionary change in mathematics, it is immature, and future developments will be constrained by inherent difficulties with proof assistants and formalization. but it's not hard to imagine it maturing over the next 30 or so years just as numerics did
1
12
358
how much oil is left. does anyone know
269
562
21,091
378,670
can ai make oil yet
10
59
2,463
38,781
a victorian gentlemen awakens with a cold sweat after witnessing a terrible vision. in his dream he saw a london populated entirely by penny-farthings, man himself being replaced just the same as the horses he once rode
2
16
824
in 6 months everyone will believe everything. the average person will understand all things to be true simultaneously. there will no longer be any propositional, linguistic, or intuitive falsity
4
34
811
the discourse around space datacenters feels intellectually dishonest to me. no matter how you set up or tear down goalposts for it, the concept is still stupid. it can only be justified with stupid premises and used to achieve stupid ends
Aaron Burnett
@aaronburnett
Sep 14
It’s possible they just don’t say it to my face anymore, but I think most have moved the goalposts from impossible to impractical and not economic within 10 years. The biggest debate left is when, Elon will almost certainly put up early versions 27/28, then the goal post will move to. “But clearly we meant scaled orbital compute”. This is a very common theme with technical innovation and anyone willing to say what they are building before it’s real. The alpha here IMO is picking the time of when StarMind will place its first 10GWs in orbit. (Or well on its way). Adding every incremental 10GWs is 300B+ revenue conservatively. If that happens before 2035 I’d wager 98% of investors don’t have that in priced in.
5
29
2,154
of course you can put an nvidia gpu in space. you can't make one giant vehicle, but you can certainly deploy a constellation of thousands of smaller vehicles. you can run inference on them, presuming you can eat tens of billions in annual operational costs. but why? why do that?
3
13
427
do you really think creating the single most complex bit of infrastructure in human history is appropriately justified by zoning restrictions in tennessee? are you so incapable of public relations you need to ascend to the heavens to escape the tyranny of a city councillor?
12
226
Load more