efficientsticks

Ok, only in iShares Exponential Technologies ETF as the 3rd entry.

There has been press hype linking Palantir with AI however:

FT: AI has given Palantir its mystique back

The Register: Former NHS AI leader joins US spy-tech firm Palantir

It’s fairer to say it probably aspires to be seen as an AI stock.

Another AI article on the front page since an hour previously:

https://news.ycombinator.com/item?id=36271120

Friendly FYI that Palantir is considered an AI stock, because unfortunately BBC and newsnotfound felt no need to mention it.

“Then, automatically file a patent for it”

Thank you for the article. During lockdown I made a search engine that you might like to try: https://highbrow.se/

No special algorithms and just the plain slightly spammy but endearing internet :)

They use a lot of machine learning for ads and YouTube recommendations - the TPU makes sense there and if anything shows how hard they try to keep costs down. It’s a no-brainer for them to have tried keeping Search as high-margin as possible for as long as possible.

Google went slow not due to ethics but because running neural inference is a lot more expensive than serving SERP data from cache.

Normal people are not rational agents.

Lots of people prefer comfort above and beyond the necessity: private cars instead of bus, takeaway instead of home cooked meals, country homes instead of small city apartments.

I agree it’s worth looking at the history, and to not repeat its mistakes, though at the same time this is a new situation, and it will continue to be new into the future, so sticking to heuristics may not serve humanity as well than being open-minded on the policy front.

The ambiguity of “advanced socialism” is problematic for any meaningful debate, so I apologise for that.

I was meaning something closer to “we have the resources and technology (in this advanced era), just not the wisdom or political will”. The actual nature of what could be provided is up for debate, but if we’re looking at mass unemployment in 2 decades’ time, perhaps it’s a conversation worth having again.

I agree, listening to the podcast I think the answer is that “yes” that is it: faith in technological progress is the axiom and the conclusion. Joined by other key concepts like compound growth, the thinking isn’t deep and the rest is execution. Treatment of the concept of ‘a-self’ in the podcast was basically just nihilistic weak sauce.

They’re dystopian fictions, ie. examples of what not to do. But experience has shown that the real-world often recreates dystopian visions by example.

So trying to be the first to show something in a well-meaning way can nonetheless have unfortunate consequences once the example is copied.

> The goal, the company said, was to avoid a race toward building dangerous AI systems fueled by competition and instead prioritize the safety of humanity.

> “You want to be there first and you want to be setting the norms,” he said. “That’s part of the reason why speed is a moral and ethical thing here.”

Clearly having either not learned or ignored the lessons from Black Mirror and 1984, which is that others will copy and emulate the progress.

The fact is that capitalism is no safe place to develop advanced capabilities. We have the capability for advanced socialism, just not the wisdom or political will.

(I’ll answer the anonymous downvote: Altman has advocated giving equity as UBI solution. It’s a well-meaning technocratic idea to distribute ownership, but it ignores human psychology and that this idea has already been attempted in practice in 1990s Russia, with unfavourable, obvious outcomes).

Good idea. A couple more non-AI engines: https://search.marginalia.nu/ and my own https://highbrow.se/

Or they are maladapting to pain that they’ve experienced, or just behaving out of pure ignorance.

It’s way past religion and it’s more like am explosion of science. GPT3.5-turbo has an IQ.. What will GPT 4’s IQ be?

Even so, it’s escaped the lab already. Which model will make the better products?

When performing calibrations I typically found my (accurate) blood finger prick monitor to be 1 mmol/L lower.

It’s not enough of a difference to be deadly, as the sibling comment suggests, and the clinician guidance is to always eat something if you are feeling hypo, and at or near that level.

The upshot is that I set my CGM-linked pump to target 1 mmol/L higher to compensate.

Type 1 diabetic here, they’re not close to as invasive as insulin pumps and a lot of people use those. Honestly the real problem with CGMs is they aren’t accurate enough, and they bias high - I suspect so they can pretend they eliminate hypos more than they do. (Having said that, CGMs are well worth it even with those drawbacks. They’re only invasive on the initial application.)

But I do applaud the team for working on the technology nonetheless.

That’s a very engaging explanation. I’d recommend Guillermo del Toro's Pinocchio as a similarly engaging way to introduce life and death early on.

Perhaps a couple more analogies would be helpful for how animals and trees can be alive too.

Thanks for this great reply.

Ironic thing though - that 10% most severe poverty mostly lies on the African continent, exactly the place which could stand to benefit the most from China’s intervening, and of course before that it was Chinese people who were in great poverty.

Now I’m as aghast at the militant astro-turfing by that country on these forums as anyone else (and much worse inside), but there I’m not so pessimistic, perhaps seeing their own positive change effected on an outside group can cause some kind of re-evaluation for the culture.

I agree that things should be considered ultimately as political (even judicial if needed).

But it’s not a given - Black Mirror describes worlds in which that reality does not win a survival-of-the-fittest competition against harsher technological regimes. It seems every time we describe these dystopias they become true, which for me is a reflection that it’s easier to destroy than to create, absent a learned culture.

Does LaMDA have sentient-nature?

This is incorrect.

Dogs don’t just survive, they thrive. In Romania the government resorted to mass killing of dogs because they bred so successfully. Also in poor countries and remote places where you wouldn’t expect there to be excess food for strays.

Dog ownership is very often slavery. The nature of the slavery is very explicit - the dog’s work is to sniff drugs all their life, or retrieve dead game, or guide people. In other cases you can read in this thread, a past owner has had a captive dog and abused them for who knows what reason.

I haven’t used it in years, but magit is really great.

Nowadays I just do a few rounds of git add -p; commit, or trusty ole git commit -av

In a past job, minimising Pull Requests was a key motivator - you can touch a larger cross section of codebase with a single branch.

It’s useful when lots of things that are genuinely distinct happen to have similar config, for example.

In an ideal world, a class would be all public fulfilling some abstract interface which is pertinent to the logical substance of the solution, and those methods would perform their operation using other components which are private final and supplied as constructor arguments.

This is called “over-engineering” by some, but really it is the most testable and refactorable approach to creating programs (other than being a team of PhD Haskell programmers).

The problem is simply that programming hours are expensive and we don’t got time for this methodology, in general, just as we don’t have time for Ada or formal methods, in general.

AI co-programming will change all of this.

Edit: this approach is different from Option 1 in the article because it’s about refactoring and not just flipping method visibilities even where it makes no sense for the interface.