verdverm

5 ms·

91.
by verdverm 22d ago
related / additional https://contraptions.venkateshrao.com/p/ea-safety
92.
by verdverm 22d ago
probably if you don't program already, unlikely if you do, they are more similar than different once you know the three classes (imperative, functional, logical) I suspect an amount of the same in human language, where learning french
93.
by verdverm 22d ago
paper the reddit OP "published" (their words on reddit) to arxiv (before they put the vouching process in place). It's what you expect if you click through. https://arxiv.org/pdf/2503.23303 Does not appe
94.
by verdverm 22d ago
It will be interesting to see how Ai helps us understand more about what goes on in animal minds through their verbal/body languages, like the mirror test for self recognition https://www.petmd.com/cat/general-heal
95.
by verdverm 22d ago
My shower thought for the day was that we should send Dario and Altman to detention (rich man's prison) for 30 days over the "hacking" incidents (negligence).
96.
by verdverm 23d ago
rich, coming from the guy who privatized the Patriot Act into Palantir
97.
by verdverm 23d ago
tide pod challenge, the "hacking" version
98.
by verdverm 23d ago
I suppose we are in (an ai) doomerism media (hype cycle)
99.
by verdverm 23d ago
paper the reddit OP "published" to arxiv (before they put the vouching process in place, it's what you'll expect if you click through) https://arxiv.org/pdf/2503.23303 Does not appear to be like wha
100.
by verdverm 23d ago
humans in the loop caught this, albeit way later than they ought to have 1. should never have reached the point of planes in the air 2. my smell test thinks the CCP is not so ignorant/arrogant to ship nuclear components to Iran right n
101.
by verdverm 23d ago
Cats might have more to say than you realize https://www.all-creatures.org/humor/dogcatdiary.html
102.
by verdverm 23d ago
love the "for ai" page, as a human https://raahelbaig.com/for-ai submitted page is good stuff too
103.
by verdverm 23d ago
While I largely agree, I suspect languages which require more memory management awareness, especially when they have opt out options like `unsafe`, will remain more challenging for the models. They have a tendency to turn things off to &quo
104.
by verdverm 23d ago
for relative comparison, google reported they process 3.2Q tokens / month this last i/o https://chatforest.com/builders-log/google-3-2-quadrillion-t... OAI reported 15B / min around the same time, which
105.
by verdverm 23d ago
agree, Trump is largely a puppet and easily manipulated, though I'd put more weight on Heritage, Bannon, and Miller, placing Vance closer to the useful idiot group
106.
by verdverm 23d ago
debatable, it's trivial to instruct them otherwise I personally tell them to use a regular hyphen in my global AGENTS.md because em dashes annoy me there are plenty of "washers" out there muddying the water
107.
by verdverm 23d ago
https://rickandmorty.fandom.com/wiki/Gwendolyn he's on a mission to save humanity from itself /s
108.
by verdverm 23d ago
1. did not call him "amazing" 2. we are "amazed" on a nearly daily basis, better not to desensitize, that's what he wants, overton window and such
109.
by verdverm 23d ago
Ironically, China has more Ai specific regulations than the US does already. Notably, they require algo applications (gen ai or not) to be registered with the government, much like all businesses need to register with the government. More r
110.
by verdverm 23d ago
It's amazing how much of a hypocrite he can be https://thehill.com/business/5603478-trump-organization-fore... (Trump org requests record level foreign worker permits, 2025)
111.
by verdverm 23d ago
it's about more than a single model being available to the public, see my peer comment written at the same time
112.
by verdverm 23d ago
They are doing fine as a shovel salesman (GPUs/TPUs, being the inference hardware provider, though I now wonder if their token processing stats are their own or combined with those who are running on their hardware too, I've been
113.
by verdverm 23d ago
The ROI is for we the users, not the companies, only issues is that the agent's attention is way easier to hijack, so you have to maintain a different kind of attention when using them and never blindly trust their outputs.
114.
by verdverm 23d ago
I'm not sure which version is the "unsloppified" one, perhaps it's derived from https://github.com/AlexSchmalex97/The-Nothing-Button
115.
by verdverm 23d ago
The third law from https://susam.net/inverse-laws-of-robotics.html
116.
by verdverm 23d ago
some history https://www.militarytimes.com/opinion/commentary/2026/01/14/...
117.
by verdverm 23d ago
Google is flailing right now, probably more a PR effort to say they are still relevant
118.
by verdverm 23d ago
https://data.opm.gov/explore-data/analytics/workforce-size-a... The DOGE train wreck is well known, lesser known is the impacts on the intelligence community, putting the nation at some of the greatest risk since
119.
by verdverm 23d ago
yup, exactly!
120.
by verdverm 23d ago
they're all using the same vendor, re: incestuous EA cabal

More ›