apsec112

21 ms·

241.
by apsec112 2y ago
facepalm Ugh, this is a horrible article. 45% of people found housing, but there was zero difference between the treatment group (given $1,000/month), a group given a large lump sum, and a control group given $50 a month! https:&#x2
242.
SF Deep Tech Week(sf-deep-tech-week.com)
1 points by apsec112 2y ago|0 comments
243.
Refocusing the CDC after Covid-19(paragoninstitute.org)
2 points by apsec112 2y ago|2 comments
244.
Mira Murati: Next GPT coming out in "a year and a half"(timesofindia.indiatimes.com)
1 points by apsec112 2y ago|1 comments
245.
by apsec112 2y ago
US taxes as a percent of GDP are close to post-WWII lows https://www.crfb.org/blogs/2023-revenue-plunge-confirms-2022...
246.
Sam Altman is not on YC's board. So why claim to be its chair?(bizjournals.com)
285 points by apsec112 2y ago|168 comments
247.
OpenAI ex-employees who work at competitors are excluded from tender offers(cnbc.com)
4 points by apsec112 2y ago|2 comments
248.
by apsec112 2y ago
How is working at NIST not "going to DC and working for the people at government wages"? He is, in fact, currently working at NIST, he's on the website: https://www.nist.gov/people/paul-christiano The Al
249.
by apsec112 2y ago
Paul Christiano has: https://en.wikipedia.org/wiki/Paul_Christiano_(researcher)
250.
by apsec112 2y ago
Daniel Kokotajlo quit OpenAI earlier this year, and was one of the only people who refused to sign their secret non-disparagement agreement, even though (at the time) OpenAI told him this meant he would forfeit his equity. The idea that som
251.
by apsec112 2y ago
Since 2016, real (inflation-adjusted) wages have risen quickly for people in the bottom quartile, while for upper-income workers they've been stagnant. Obviously, most people on HN are still better off than manual laborers, but the gap
252.
by apsec112 2y ago
AI has seen rapid efficiency improvements every year for the last decade, but the amount of computation used has gone up, not down. The improvements always go into making larger and more powerful models, not into just running the same model
253.
Microsoft admits whistleblower GPT-4 claims, reversing previous denial(twitter.com)
6 points by apsec112 2y ago|0 comments
254.
by apsec112 2y ago
People don't divide cleanly into "good" and "evil" buckets, and CEOs in general tend to be ruthless deal -makers. But it's worth noting that much of Sam Altman's presentation is just a mask (one he puts a
255.
by apsec112 2y ago
Prior to vaccines, there was no difference in COVID death rates between red and blue states https://www.natesilver.net/p/the-2-key-facts-about-us-covid-...
256.
by apsec112 2y ago
This page is a summary of the testimony which is clearly trying to pick quotes and/or frame things to make Fauci and Democrats look bad. FWIW I agree that eg. children shouldn't have been required to wear masks in school, but I do
257.
by apsec112 2y ago
The response to COVID should be thoroughly investigated, but this page just seems like a partisan hackjob. The framing is clearly intended to get maximum partisan impact, not to find the truth.
258.
by apsec112 2y ago
This is a really biased and misleading summary. In particular, it's vanishingly unlikely that anyone will have criminal penalties applied. More analysis here, going over the entire text of the bill in detail: https://thezvi.
259.
by apsec112 2y ago
So the idea is that what Ilya Sutskever does is "meetings and political maneuvering", while what Sam Altman does is "delivering results"? That sounds rather strange....
260.
by apsec112 2y ago
It was probably flagged because it's a dupe (both Vox stories were already covered on HN)
261.
The Myth of Panic (2021)(palladiummag.com)
39 points by apsec112 2y ago|50 comments
262.
by apsec112 2y ago
There's a full length article here: https://www.vox.com/future-perfect/2024/5/17/24158478/openai...
263.
by apsec112 2y ago
()
264.
by apsec112 2y ago
It's not. The earlier tweets explain: the initial agreement says the employee must sign a "general release" or forfeit the equity, and then the general release they are asked to sign includes a lifetime no-criticism clause.
265.
by apsec112 2y ago
The Vox article says that it's a lifetime agreement: https://www.vox.com/future-perfect/2024/5/17/24158478/openai...
266.
by apsec112 2y ago
Non-disparagement clauses forbid all negative statements, whether true or not. https://www.clarkhill.com/news-events/news/the-importance-of...
267.
Ex-OpenAI staff must sign lifetime no-criticism contract or forfeit all equity(x.com)
449 points by apsec112 2y ago|5 comments
268.
by apsec112 2y ago
This honestly feels kind of silly. Back-of-the-envelope, this training run (13B model on 380B tokens) could be done in two months on a single 8x H100 node using off-the-shelf software, at a cost of around $35K from a cloud compute provider.
269.
by apsec112 2y ago
Interesting! Would this support fp8? Does anyone know how it would compare to Triton?
270.
by apsec112 2y ago
Those are really sweeping conclusions, considering the experiment is just a single iteration of a single prompt! FWIW, Claude Opus got this for me on the first try: "In the Gilligan's Island episode "Seer Gilligan" (seas

More ›