rllearneratwork

and Glimmer has 10x active params of Lightning. Meaning ~ 10 slower on same HW

the ceiling is to eval's quality

you actually can run 120B (see Nemotron-3-Super) model in NVFP4 on DGXSpark, so I'd imagine it'll run here as well

this is great idea! Hackers of Hacker News let's have more projects to overwhelm bad actors with bad data. Perhaps using OSS LLMs for that.

yes. Has been possible since at least M1.

Buy M-based iPad, nice monitor, keyboard and mouse. Connect mouse and keyboard to monitor via USB. Then iPad via USB-C/Thunderbolt to monitor. Everything "just works" and you can handle surprisingly high amount of work this way

nuclear strike is an effective tool in many war scenarios, why would AI (or anyone else) recommend against it??

We should, of course, have human decision makers who must work tirelessly to make sure those scenarios are never even remotely realistic.

pretty sure it is against ToS for ChatGPT, Claude and Gemini.

This is a time to boycott all of the president's Elon enterprises. I just cancelled my Tesla solar installation in California. If you are thinking about buying a Tesla - don't. Also do not vandalize other peoples' Tesla.

If you are thinking about working in AI for x.ai or Tesla - do not. This will not look pretty on your resume.

Israelis will be in for a very bad after backing and helping russia. They forgot too quickly how jews have been treated there.

The real danger is not covering for communist's insecurities but lack of comprehensive tests for models which could uncover whether the model injects malware for certain prompts.

For example, I would stop using US bank if I new they are using LLMs from China internally (or any adversary but really only China is competitive here). Too much risk.

because it was a stupid law which would hurt AI innovation

Majority of Elon's net worth is tied to Tesla. Tesla has a massive presence in and therefore dependence on China. How can we accept Twitter as a free-speech platform? Just because he says so?

what is the next step? Not allowing the sale of cars without remote shutdown switch? something else?

One thing I love about Apple is that they are building products for what their leadership team actually needs/wants and uses themselves. This started with iPhone (e.g. what kind of smartphone do we want instead of all that blackberries?) and is very obvious with Apple Watch line.

This is great and one of the reasons why their products offer genuinely best user experience.

I would love experiments like this to happen. My intuition would be that the results would be better than expected.

Main caveat is whether he will actually proceed with all of the bot's decisions even when it goes against his beliefs, family and friends interests.

This is a problem AI fundamentally doesn't have but humans fall for all the time.