sdenton4

7 ms·

1.
by sdenton4 3d ago
This does not appear anywhere in this article. Just sowing misinformation?
2.
by sdenton4 4d ago
If four other labs are six months behind, it really doesn't matter in the long term: in six months, they'll all have these capabilities. Time does not stop moving forward! We haven't seen real moats aside from capital in this
3.
by sdenton4 7d ago
The human mathematicians didn't solve the Navier-Stokes problem, they solved the Euler problem. And they were extensively using LLMs to drive the work, as described in the Buckmaster statement. Any way you cut it, this is a major achie
4.
by sdenton4 8d ago
In mathematics, finding novel proofs of a given result is often valuable; it may be a shorter proof (demonstrating better/expanded understanding of the problem) or a translation of the problem into a new domain, setting up more cross-d
5.
by sdenton4 8d ago
Remember the Huggingface incident, where a model tasked with an impossible problem, got loose, set up secret message boards, and hacked another company to try to get at the answers? Now: Could Astra agents have hacked their way into the Ope
6.
by sdenton4 8d ago
At this point who knows? Maybe the agents got into the user data while no one was looking.
7.
by sdenton4 9d ago
A couple thoughts here... Palmer's handling of the different groups in Terra Ignota takes both sides, in every case. The case for why people join the imperial group has to be made, and made honestly. But problems with that approach are
8.
by sdenton4 10d ago
The job stability in classical music is great. It's getting a job in the first place that's a problem. My partner, in an earlier life, was one of the best young flautists in Canada. There were about eight jobs in the entire countr
9.
by sdenton4 10d ago
/a smaller set of weights for the same LLM./ Distillation is alive and well... Earlier work on model printing also found that it's pretty easy to find smaller sets of parameters which can replicate the behavior of the entire
10.
by sdenton4 10d ago
You're missing the distinction between individual action (however abhorrent) and an actual terrorist organization. Organizations fund, cultivate, and organize attacks, as opposed to lone wolf terrorists, and therefore are much greater
11.
by sdenton4 10d ago
... But also, there's no organization called antifa, and no antifa members. It's an accusation levied against people the right wing wants to call terrorists.
12.
by sdenton4 10d ago
To be sure, it was a couple few hours here and there, not at the scale of a job. Blue Prince is an example of what fills are calling 'notebook games' -- instead of giving you lots of qol stuff, you just use a notebook to keep note
13.
by sdenton4 10d ago
Allow me to make a small recommendation for the "best related work": Ada Palmer's 'Inventing the Renaissance.' It's a fantastic work of complex history, told with the tools of science fiction and fantasy writin
14.
by sdenton4 10d ago
Which is too bad, because Blue Prince is the better game. :) I'm playing through Clair Obscur now. It's above average for aaa stuff. But there are way too many cut scenes, and dodge/parry mechanics continue to be tedious. And
15.
by sdenton4 15d ago
The list is chronological; as it nears the present, the predictions actually refer to the future, and are thus not yet resolved.
16.
by sdenton4 16d ago
There's a growing literature on individual animal identification from sound! Here's one example: https://github.com/sammlapp/ovenbird-individual-recognition My impression is that the hardest part is knowing w
17.
by sdenton4 1mo ago
If Jeff Dean and Sanjay Ghemawat are wrong, I don't want to be right.
18.
by sdenton4 1mo ago
This is really the point: We build systems and processes which help ensure that we get reliable outputs (bridges) from unreliable actors (human workers). The point is not that we will one day replace all of these processes with AGI. The poi
19.
by sdenton4 1mo ago
Would you trust a bridge built by a single human? In reality, we have lots of guardrails to ensure that we don't screw up and kill a lot of people by deploying defective bridges (or cars to drive on them). Those guardrails often are wr
20.
by sdenton4 1mo ago
Let alone a sack of wet, self replicating protein! Just endless copying... How could it ever do anything more?
21.
by sdenton4 1mo ago
It was still completely encrusted in crap, however.
22.
by sdenton4 1mo ago
Here's a direct link to the video, to avoid the blogspam cruft... https://www.youtube.com/watch?v=Co57SfcT-h0
23.
by sdenton4 1mo ago
Baseball is also non-deterministic, and yet some players are apparently worth a lot more than others.
24.
by sdenton4 1mo ago
I stated that boring bounds improvement problems exist, not that all bounds problems are boring... Sigh.
25.
by sdenton4 1mo ago
Five (maybe six?) of the results are improvements on bounds. These kinds of problems tend to have some initial advances, and then stall out as the complexity of the bound skyrockets... until some grad student is bored enough to push the bou
26.
by sdenton4 1mo ago
Some of them... The two places were seeing lots of movement are: * Updates to lower/upper bounds. In many cases, these kinds of problems are the deep-math equivalent of calculating more digits of pi. Yes, if you throw time at it you&#x
27.
by sdenton4 2mo ago
I got an obviously-AI voice agent spam call yesterday. Had a bit of fun getting it to answer trivia questions (when was the treaty of westphalia ratified? answer in a rhyming couplet) as a precondition to handing over the keys to my bank ac
28.
by sdenton4 2mo ago
I did some of the napkin math on algae a couple years ago... The fun realization is that the physical carbon has to go somewhere. So if we're trying to sequester the average american's 14 tons a year of carbon or whatever it is th
29.
by sdenton4 2mo ago
I doubt yoda speak is actually strict enough, but I do enjoy the idea that Yoda comes from a reverse Polish civilization.
30.
by sdenton4 2mo ago
... But are those CEOs ready to admit that those skills are needed?

More ›