matteoraso

>We never stop learning, we don't have a "training phase".

We kind of do, if you count critical periods in childhood.

But the EU was never meant to be a loose forum of nations. From the very beginning, the Single European Act made it clear that the intention was to integrate the member states into a single union, "speaking ever increasingly with one voice".[0]

[0]https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELE...

>The concern is that autism as a rigorous category may be useless, but that doesn't mean that autistic people are suddenly neurotypical

IDK. I'm diagnosed as autistic, but I've become a bit less attached to that label in recent years. Like, I do struggle with eye contact, excessive noise, and social situations. I'm not going to say that I'm just a normal guy, but I do think that there's a wide spectrum here that encompasses everybody. I don't buy the idea that autism is some special category of people that are totally unlike neurotypicals. Instead, it's just one side of the spectrum, with neurotypicals on the other end, and an ambiguous group in the middle[0]. I think that once society becomes more accepting of neurodiversity, we're going to stop getting so hung up on labels like "autistic" and "neurotypical" and accept that there's a lot of nuance in the human condition.

[0] In reality, this spectrum would be multi-dimensional, but I'm choosing to describe it as a 1D line for illustrative purposes.

This is true, and an understated consequence of homelessness. My city is working on deploying portapotties for the homeless for this reason.

It's also overseen by the Minister of State at the Department of Health and Social Care. Also, how on Earth do you think insurance companies don't consult doctors on insurance denials?

Something that gets ignored is that we're already facing a serious decline in the workforce as a result of the fertility crisis. If we can automate some jobs at the same time as people leave the workforce en masse, we can avoid a catasrophe.

>I think a lot of people would be very content if they never got smarter, and just kept getting even cheaper/faster.

There's a lot of truth to this. I think we're starting to approach the point where increased intelligence has declining marginal returns, such that it might not even be worthwhile to improve models unless it can be done cheaply.

The world economy is based on the idea that America will always repay its debt. Them not paying it back will be the financial equivilent of a nuclear bomb, not just for them, but for literally everybody.

The uncomfortable reality is that fixing this will require tax hikes AND spending cuts. Not just taxes on people you don't like and spending cuts on niche programs, real changes like a VAT and slashing the VA.

Often the case with big pharma. I for one will never join in on the populist backlash against those that make livesaving pharmaceuticals for us.

Huh? How would chemo "block his liver"? Do you mean that he had a liver metastasis that the chemo didn't help with?

Huh? NICE sets a cost-effectiveness threshold of about £20,000-30,000 per QALY gained from a drug[0]. There was even a controversy where a drug for pancreatic cancer was denied funding for this very reason[1].

[0]https://web.archive.org/web/20110925050527/http://www.nice.o...

[1]https://web.archive.org/web/20151225115559/http://www.pancre...

>With Universal Healthcare the government just provides the funding.

And what makes you think that they fund everything? I support universal healthcare as well (I'm literally Canadian), but I can accept that there are tradeoffs associated with it.

The problem is that LLMs improve so fast, your article may very well be out of date by now.

RAM consumption has gotten crazy over the past couple of years. Consumers need to start calling software devs out for this, especially when there's such a severe shortage of RAM.

I don't see what this offers that Git doesn't. Even if Git only tracks lines, that's still a good proxy of how many words you wrote. Besides, you can use git diff to see whether an individual character was changed, even if the number of lines stayed the same.

You're judging a singular product and then assuming that this is all the foundation does. If you want to know why they have so many employees, you have to look at the scale of the entire foundation, not just one of their websites.

So you're judging the entirety of the Wikimedia Foundation based on the UI of one of their websites? This is an insane way to measure an organization.

You're unfamiliar with the scale of Wikimedia. Besides Wikipedia, they also host multiple other projects, like Wikidata and Wikimedia Commons. Besides that, a huge portion of their budget goes towards grants to organizations that help distribute information for free on the internet. This absolutely isn't a "CRUD PHP site".

It's simple economics. Building a robot to do your chores is expensive and only a small minority of people value their time enough to buy one. Meanwhile, SWEs are expensive and GPUs are (comparatively) cheap.

People have had to deal with getting their jobs automated away for centuries. None of this is new, and perhaps reminding ourselves of this is the best way to cope.

>A pharmacy chain in Vermont implemented AI for efficiency. It’s led to delays, incorrect information and privacy concerns.

You should post the entire title.

>Have LLMs produced something novel like Cantor's diagonalization procedure once was?

A year ago, the goal post was that LLMs wouldn't prove anything of worth. The year before that, it was that LLMs wouldn't be able to do math as well as a undergraduate student. Year before that, it was that LLMs wouldn't be able to do math at all. The growth of LLMs has been insane, and it doesn't look like it's going to plateau any time soon.

>2. OpenAI is clearly incapable of controlling their in-house models. This is the second time Galaxy-class models are known to have breached containment and done bad stuff.

What was the first?

Then it sounds like your energy bill wasn't because of a data centre.

I've been wanting to experiment with using AI to prove math theorems, but compute is obviously a massive limiting factor here. Are there any plans to open source this?

They've already attacked Syria and Lebanon multiple times already. There's no doubt that they're already planning for new wars once this one is over.

I'm in the same boat as you. Even if there's slight issues with Firefox, being able to synchronize my profile with my phone using the Firefox app outweighs all of that. AFAIK, Chrome doesn't have that.

Maybe not exactly what you're looking for, but I highly recommend "The Real North Korea" by Andrei Lankov. It really helped to demystify the country for me.

Check out Libreoffice. It's like MS Office, but less resource-intensive.