doctorpangloss

6 ms·

1.
by doctorpangloss 4d ago
> For example, there is a 100% test coverage requirement in both server and client, combined with AI-driven review rules that say all tests must be non-vacuous, look, your application works, right? so it doesn't really matter what y
2.
by doctorpangloss 4d ago
the requests went into a pipeline that turns them into de-identified, but salient, training data, yeah. everywhere except maybe bedrock.
3.
by doctorpangloss 5d ago
It's complicated. SMBs can find an audience that would otherwise be impossible to reach. The thing that is advertised is so diverse it's hard to generalize, however the main thing I've noticed is SMBs live and die by the qual
4.
by doctorpangloss 7d ago
I don't know. Apple marketing has been really effective at this health thing. Whereas, studies of continuous monitoring of eg heart with even better sensors and higher risk populations have shown conclusively no improvement in health o
5.
by doctorpangloss 8d ago
> If they opted out of training, then we definitely did not train on them. are opted-out-of-training chats ever paraphrased, and thus "de-identified" (in openai's own words)? what this means is, it would be hard to prove t
6.
by doctorpangloss 8d ago
> It would be extraordinarily easy to simply say, this model was not trained on your work, if that were the case. well, it is trained on their work. all user inputs are paraphrased for training. at openai, at anthropic, at google, and no
7.
by doctorpangloss 8d ago
> While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models it just means that they've been working with paraphrased user data everywhere, which any smart person ca
8.
by doctorpangloss 10d ago
the reality is that those stats were flawed in numerous ways. this is expected, they were never subject to peer review. for example: ok cupid and tinder, during the times that christian rudder was blogging about them, never reported on the
9.
Mother convicted of: her 5 yo walks short way to pond alone in Virginia(washingtonpost.com)
26 points by doctorpangloss 10d ago|21 comments
10.
by doctorpangloss 14d ago
they basically map out the paying audience for your game on Steam
11.
by doctorpangloss 14d ago
The most valuable data is the per user purchase histories from before the time those were hidden by default.
12.
by doctorpangloss 14d ago
i don't know. the whole operation was kind of obnoxious. valve could provide all this information for free in an afternoon, but chooses not to. of course, by information, i mean: historical, per-user purchasing histories, which are 1,0
13.
by doctorpangloss 15d ago
Indeed.
14.
by doctorpangloss 15d ago
"Can reconstruct [scenes from Unreal Engine]"
15.
by doctorpangloss 16d ago
For mathematicians in the US, it's because their parents like money.
16.
by doctorpangloss 20d ago
Deepfakes are kind of the point though.
17.
by doctorpangloss 20d ago
[left hand emoji] you have very strong feelings about something, you're sharing it with others on a public forum [right hand emoji] "their political stance"... "they should keep them to themselves"
18.
by doctorpangloss 20d ago
on the flip side retro games are almost completely and entirely an aesthetic experience. like if something totally subjective and personal like nostalgia is not just a valid reason, but the #1 reason, to be engaged with retro gaming, why is
19.
by doctorpangloss 21d ago
for example, their people think that GB300s are twice as fast as B300s, when really their benchmarks just incorrectly divide GB300 instances on azure by 8 instead of 4, since they don't read or verify any of the code that executes thei
20.
by doctorpangloss 21d ago
personally i reject any frame of mind that makes stubborn, labeled generalizations about kids based on one or a handful of interactions. the people who conduct their business this way: they are always projecting.
21.
by doctorpangloss 21d ago
when florida banned phones from schools, a gold-standard causation study found a benefit of 0.9 percentiles on standardized tests between students who were under the ban and those who could use phones freely. this is about the same impact a
22.
by doctorpangloss 21d ago
it may be possible for some people, I don't know. does Meta have internal research that uses the word addiction that would come up in discovery? probably. so there's more than one side of this. a little bit of a chance in the heal
23.
by doctorpangloss 22d ago
i'm aware of this one. clinical evidence - do moderately high CO2 levels CAUSE a disease? afaik no. the test in this study: they measure something real, but it is not used to diagnose someone routinely (despite what they say). they had
24.
by doctorpangloss 22d ago
The semianalysis people have scripts which incorrectly count their numerators and denominators all the time. All their benchmarks are flawed. It is such a slipshod operation and they charge exorbitant amounts of money for it.
25.
by doctorpangloss 22d ago
There isn't really any clinical evidence for CO2 levels impacting stuff beyond so called "sleep efficiency," where the effect is small (above 1000 ppm is 5% reduction in sleep efficiency) and it has the counterintuitive effec
26.
by doctorpangloss 22d ago
Legitimately a much better game than actual Wordle.
27.
by doctorpangloss 23d ago
there are a lot of completely free file sharing services for authentic activity (e.g. github, google drive) inauthentic activity seems to move from one product space to another. in 3 years we'll be hearing about stripe shutting down op
28.
by doctorpangloss 24d ago
hardest for claude and chatgpt, which will refuse to do this task haha
29.
Suspicions of hitting daughter between a congressman and his ex-wife(washingtonpost.com)
4 points by doctorpangloss 26d ago|1 comments
30.
by doctorpangloss 26d ago
i'm not saying they reflect how much you study. they reflect a lot of things, including that. but you ask the people who write the tests, they're going to say, how much you know or how much you study, but nonetheless, they are lim

More ›