mentalgear

25 ms·

31.
by mentalgear 20d ago
Title should be : "We monitor internal coding agents for misalignment, we really do, pinky swear." Actions are louder than words.
32.
by mentalgear 21d ago
Plus DuPont Chemicals dumped forever chemicals like PFOS directly into swimming lakes, rivers, and the ground where it also ended up in communities ground water.
33.
by mentalgear 21d ago
There's an excellent film with, and produced by, Mark Ruffalo ("Avengers"). "Dark Waters" about the environmental lawyer who exposed the whole fowl conspiracy by DuPont Chemicals, 3M, and other industrial chemical c
34.
by mentalgear 22d ago
I'm always wondering whether those centralized privacy services are not the easiest first target for three-letter-agencies to infiltrate to gain access to the most relevant users to track - and what currently would prevent them from do
35.
by mentalgear 22d ago
their marketing is really low-key and they must improve it because they have enticing products. OS offers page: https://www.shift.eco/shiftos/
36.
by mentalgear 22d ago
Wouldn't be surprised: Snowden's revelations 10 plus years ago already showed how the NSA was injected into the data centers of Social Media, it's only logical that they would now demand to be injected into the biggest, most
37.
by mentalgear 22d ago
So OpenAI’s stance on AI safety is now basically that Blues Brothers meme: two guys in dark sunglasses, driving at night in a car with broken headlights, pedal to the metal, asking, "What could possibly go wrong ?"
38.
by mentalgear 22d ago
It's called 'pay-for-play' corruption, aka the only leading principle of the current US admin.
39.
by mentalgear 22d ago
Great what they are doing - yet I found Shiftphone to be also fair, equally or better repairable and even more open for alternative Operating Systems. I think Fairphone supports 2 OSes officially, Shiftphone has a whole custom ROM list of ~
40.
by mentalgear 22d ago
So OpenAI’s stance on safety is now basically that Blues Brothers meme: two guys in dark sunglasses, driving at night in a car with a broken windshield, pedal to the metal, asking, "What could go wrong ?"
41.
by mentalgear 22d ago
So OpenAI’s stance on interpretability (ai safety) is now basically that Blues Brothers meme: two guys in dark sunglasses, driving at night in a car with a broken windshield, pedal to the metal, asking, "What could possibly go wrong ?&
42.
by mentalgear 24d ago
With their security, they probably still don't the know the full extend what may have happened that or the last time. Might be another swarm of agents currently colluding somewhere in their sub-sub-infra - possibly striking critical in
43.
by mentalgear 24d ago
I'm looking forward to an announcement of them making Alignment Top Priority - as it should be, especially giving their alarming breach of 700 agents colluding outside of their knowledge for months culminating in hacking HF (here'
44.
by mentalgear 25d ago
> Forced tool use is not supported That seems unfortunate for 3rd party integrations that expect stable output - what that really necessary ?
45.
by mentalgear 26d ago
Everything hyped is usually a counter quality signal
46.
by mentalgear 26d ago
Friends dont let friends use Omarchy or [claw] products.
47.
by mentalgear 27d ago
Credible Source or it didn’t happen .
48.
by mentalgear 28d ago
What if it turns out the real 'AGI' was the recording of billions of 'thinking traces' from (paying) users giving feedback and guiding the model - so LLM providers could extract their thinking and privatize it ?
49.
by mentalgear 28d ago
I wish ShiftPhone - which is fitting all my other needs - would offer a config that would work for GrapheneOS.
50.
by mentalgear 28d ago
It isn't, and it is not what was described: its about unbounded imagination (neural / LLM) that needs reality constraints (symbols / rules) to produce useful output. Think of it like human imagination may do anything (flying
51.
by mentalgear 28d ago
LLMs using REPL are one instance of symbols to "bounce" their prediction against domain constraints for verification. Also shout out to Gary Marcus who was right after all (and LLM companies wasting 100s of billions of dollars for
52.
by mentalgear 28d ago
In general, what all the big LLM providers are doing is moving towards classical & neural (neuro-symbolic) AI - even though they dont publicly admit it because that would counter their claims for years of "scale is all you need&quo
53.
by mentalgear 1mo ago
Pin a post that you leave X (or any major SM): say its your last post because they have become a greedy, amoral cesspool and add links where they can find you instead.
54.
by mentalgear 1mo ago
Just do the obvious thing: stop posting there, don't view it anymore, and pin a post that you leave X (or any major SM) because they have become a greedy, amoral cesspool and links where they can find you now.
55.
by mentalgear 1mo ago
This is an awesome blast-from-the-past UI - I checked out Obama's profile and it made me almost confuse it with the real twitter from ~10 years ago. But it may only takes another X-ed lawyer letter to make it go all away again. I wonde
56.
by mentalgear 1mo ago
TechnoKing!
57.
Meeting Seed7(genodians.org)
2 points by mentalgear 1mo ago|1 comments
58.
by mentalgear 1mo ago
> It turns out that long before memory-safety had become a fashionable topic at tinder dates, the problem had been identified and attacked by a programming language called Seed7. "Long before" as in three decades ago. "Att
59.
by mentalgear 3mo ago
This echoes 'safe fracking' claims - and now many people in proximity have gas coming out of their facets. Digging hazardous materials out of sight into a potentially unstable or potentially becoming leaky structure is never a sou
60.
by mentalgear 3mo ago
This might be more about whatsapp poor organisation tools, the general need for private ML-based smart indexing/notifications and efforts at structured community building.

More ›