Sune

Why did this get so downvoted without any explanation? They have been stealing pictures, how can we be sure that have not been stealing some ideas towards the counterexample of the conjecture about Navier-Stokes?

Great quote for the Guardian article linked above:

In response to those concerns, Altman and his counterpart at Anthropic, CEO Dario Amodei, called for the industry to “pace” the development of AI […]

Even so, both companies rolled out new models on Tuesday.

Had a human carried out this attack on Hugging Face’s infrastructure, they could have faced felony charges.

Someone should make a website like https://vibemathed.com/ but with crimes committed by AI and listing minimum and maximum penalties for the crimes if it had been done by humans!

By professionalism I don’t mean wearing a suit. I mean behaving respectful, responsibly, and with integrity. Basically the opposite of trolling.

Also, I would not expect the dress code at xAI to differ from Anthropic and OpenAI. Do you think it does?

I would say part of Grok’s brand is being unserious, trolling and unprofessional, and this rubs off on the reputation of xAI. This is mostly orthogonal to left-vs-right. It is an attitude currently embraced by the American right, but historically it is an attitude the right opposed.

On top of that, Grok is of course right-wing, but it also has brand issues orthogonal to that, that could make right-wing researchers want to not work there.

When it comes to Grok, I don't feel like Elon is not hiring more researchers at xAI because nobody is willing to work for it.

The number of researchers he could hire is not affected by the brand, but the quality of researchers is.

That's not how pharma works.

I completely agree. But what matters for a cascade is how people think anti-aging pharma would work.

Im surprised by this, so I think it would be valuable to have as a top level post with more details.

You are one day old

I think your comment was meant positively, but this could come across as saying the question was naive. I think it was a very relevant question, and although I have followed the AI safe debate for a decade, I had the same question.

Who do you see as potential adversaries that would do something like 1? I can see that nation states would be intersted in infiltrating the AI security community, but not why they would want to weaken its credibility.

Rogue agents would have an interest in weakening our credibility. Do you think this is currently a threat?

Labs might have a short term interest, but it would be big risk for them to do something like this, and I don’t think they would do it. Individuals from labs might want to do this, but it would be a lot of work for an individua.

the "safety is for marketing" meme

I think part of the problem is that “safety is for marketing” can refer to two almost opposite statements. 

1. (False) AI is not dangerous, but labs plays up safety concerns to make their technology seem more powerful.
2. (True) AI is dangerous, and AI developers know this. For that reason, labs want to emphasise their work on safety to be able to attract the best employees. Sometimes labs “safety” efforts are more about appearing safe than actually doing something meaningful.

When trying to communicate 2 it is important to not do ... (read more)

How about just tapping a glass repeatedly (with either cutlery or another glass)? This would be the usual way to get the attention in Denmark at least. The sound is easily heard over conversations.

How long can the company drag out the case before they have to pay compensation? What can you do if they just keep saying that they are working on it? On a previous occasion it took me more than 3 months to get the compensation, and I have a current case that has already taken a month.

I was sceptical, so I tried to make Claude guess who wrote one of the test paragraphs, and it guessed correctly and was 75% confident. I asked it not to use internet, but i didn‘t verify that it didn’t use google.

I feel like hiring wise, this post is really very substantially about experience.

I think that wasn’t that clear from the post. I read it as saying that you can have a talent for each of the 4 skills, and that experience with each skill isn’t transferable. Past experience in a skill is only relevant as a proof that you have a talent for that category.

Why is 0.1% not being monitored?! Naively, I would expect it to only be around 0.1% extra work. Or is there a reason that the last 0.1% is somehow harder to monitor?

Does that mean that the brain itself was simulated, using only info from the brain scan, while other techniques were used for simulating the body and environment?

Was it a specific fruit fly, or a simulation created using data from multiple fruit flies?

Is there any hope/indication that this process can be reverted? For example, has this ever been reverted in a single cell? Or do you think this treatment is only relevant for uploading?

What kind of visa do you need for this? Do plans to use MAiD counts as plans to leave the country?

Assuming you don’t do this until you are terminally ill, what is the chance of having enough time to plan this, rather than needing emergency service?