sigmoid10

8 ms·

1.
by sigmoid10 3d ago
I would say in this case abliteration is the likely culprit. To uncensor a model this way, you literally deactivate the parts that would enact refusals. As in things it was told not to do. But the real process is more like brain surgery per
2.
by sigmoid10 3d ago
You'll need tons of tokens for these to churn through heavy reasoning and feedback driven tasks. You might even want to run many agents in parallel. Combine that with GPU memory requirements in the terabyte range just to hold the weigh
3.
by sigmoid10 3d ago
GLM 5.3 is probably the best open weight model for cybersecurity/exploit development right now. Though it is still significantly behind the proprietary ones and you probably need your own datacenter to run it effectively. Same goes for
4.
by sigmoid10 3d ago
This kind of argument is always dangerous, because it essentially resorts to moving goalposts. "Oh, AI can now do X? Sure it is amazing, but it can't do Y yet, so we're totally safe!" Just because humans can't proce
5.
by sigmoid10 3d ago
That feels oddly similar to the usual conservative-think that "guns don't kill people, people kill people." Yes, that is technically true. But guns make it dangerously easy for even the dumbest and mentally weakest people to
6.
by sigmoid10 4d ago
More like they were trained to complete a very specific task that has a known solution using all available tools and methods. Give an average human these levels of IT skills and tell them their future depends on the solution, they too will
7.
by sigmoid10 4d ago
Not to mention 13 em-dashes in a 500 word post. At this point you may as well ask Claude yourself about news and cut out the middleman. That'll also cut into their attempts to manipulate your opinion.
8.
by sigmoid10 6d ago
But they didn't know which problem had a possible solution. So they fired it at a huge amount of problems and then merely focused on the one that turned out to be promising. The model found the promising path itself, they just re-all
9.
by sigmoid10 6d ago
Technically it's any medium. The lower the refractive index, the closer the particle needs to travel to the speed of light in vacuum. But you can for example measure Cherenkov Radiation in the air (where n~=1.0003 or 99.97% of c) from
10.
by sigmoid10 8d ago
Especially with the current rate of progress, you asking ChatGPT today probably means you'll get a way better answer than whatever the author got using a model from last year.
11.
by sigmoid10 8d ago
That's more or less what they did according to their announcement. They fired it at a whole bunch of high end math problems and merely concentrated all efforts on one after it made some promising progress.
12.
by sigmoid10 9d ago
I actually had so many troubles with vmware on windows over the years that once I tried out hyper-v for the first time last year, I somehow have never touched vmware again. Sure vmware has better configs, cleaner interface and some very sub
13.
by sigmoid10 10d ago
>but with good repos and good prompts I think waaay more people struggle with this than HN would have you believe. In the real world, not everyone is a software dev with a developer mindset to using these tools. Normal people essentially
14.
by sigmoid10 10d ago
>Agent First’s supercluster of real, well-behaved browsers bypasses or solves reCAPTCHA, Cloudflare Turnstile, hCaptcha, & other captchas Can you share how you prevent spammers from simply using this to do what they usually do, or, i
15.
by sigmoid10 12d ago
More like not ignoring the company owned by one of the president's biggest donors warning said president's administration about your model. Anthropic basically tried to close their eyes to make this go away. Which of course backfi
16.
by sigmoid10 12d ago
If you have followed news reporting, you probably heard that SamA was touring D.C. to make sure this release went without any regulation hiccups. If anything, they learned how to play the whole politics game - especially after the Anthropic
17.
by sigmoid10 19d ago
Sorry, you are still arguing against things I never said or that you desperately want to misinterpret in a disingenuous way. If I was actually wrong about anything I said, all you had to do was write down the explicit metric and point out e
18.
by sigmoid10 19d ago
Mathematics in the sense of a complete set of axioms can't, but human research into mathematics apparently just needed enough compute to achieve the same output as a high-tier faculty.
19.
by sigmoid10 19d ago
If OpenAI has set the API standard more or less without even trying just by offering one of the most straightforward ways of interacting with LLMs via http. To the point that they got stuck with this legacy thing now that they and many othe
20.
by sigmoid10 20d ago
These days stripe is perhaps getting dominant, but there was an age before stripe and many small companies still have custom solutions. And how would Paypal scam you out of your money? I've only ever seen this happen to people who did
21.
by sigmoid10 20d ago
I meant scandinavia.
22.
by sigmoid10 20d ago
Do you never order stuff online from outside of scandinavia? I've used it extensively for over a decade with zero issues and still find it the best payment method for online shopping if you are a consumer. You only need an email to sen
23.
by sigmoid10 20d ago
USB was developed by a big alliance of major software and hardware vendors. This stuff on the other hand always feels like Anthropic is just scared of dropping out of the news cycle because other companies are constantly releasing models or
24.
by sigmoid10 20d ago
>You wrote that the singularity is a point in space Because it is. Remember: space, not spacetime. Hence the remark in brackets in the original comment and my reminder to read precisely in the other one. And in Eddington Finkelstein it i
25.
by sigmoid10 20d ago
The singularity is just a bookshelf and love can act as space-time GPS amirite? Slap it on a Möbius strip on and make it a flashy wearable, then you basically have the last Avengers movie too.
26.
by sigmoid10 20d ago
That's what you get when you start reading wikipedia about such an advanced topic without knowing anything about it. Wikipedia actually starts to crack as a reliable source for laypeople and on top of that begins to mix science and met
27.
by sigmoid10 20d ago
Huggingface is not a product, it's a community. It's the go-to place where everyone who has anything to do with open AI models gathers. You could trivially reproduce their stack, but building such a huge AI community around a sing
28.
by sigmoid10 22d ago
Just draw the geometry in Eddington Finkelstein coordinates and you will see everything I wrote above is true at the technical level if you read precisely.
29.
by sigmoid10 22d ago
Of course you can show stuff with them too. You can show tons of things with all of them. I was talking in relation to Schwarzschild coords.
30.
by sigmoid10 22d ago
It is not difficult to construct geometries with naked singularities. Reissner, Nordström, Weyl and others individually came up with one long before Oppenheimer. You can also construct geometries where faster than light travel is possible.

More ›