News for hackers. Stuff that matters.

sigmaisaletter

Re:Claude 4 on Thursday May 22, 2025 @06:33PM (#44067881)
Anthropic System Card: Claude Opus 4 & Claude Sonnet 4

Online at: https://www-cdn.anthropic.com/6be99a52cb68eb70eb9572b4cafad1...

Re:The "AI 2027" Scenario: How realistic is it? on Thursday May 22, 2025 @05:16PM (#44067144)
> 10 years ago, the entire conversation would have seemed ridiculous

Bostrom's book[1] is 11 years old. The Basilisk is 15 years old. The Singularity summit was nearly 20 years ago. And Yudkowsky was there for all of it. If you frequented LessWrong in the 2010s, most of this is very very old hat.

[1]: https://en.wikipedia.org/wiki/Superintelligence:_Paths,_Dang...

[2]: Ford (2015) "Our Fear of Artificial Intelligence", MIT Tech Review: https://www.technologyreview.com/2015/02/11/169210/our-fear-...

Re:The "AI 2027" Scenario: How realistic is it? on Thursday May 22, 2025 @05:08PM (#44067044)
I think you mean "somewhat naive" instead of "somewhat native". :)

But, yes, this, in my mind the peak[1] bubble times ended with the DeepSeek shock earlier this year, and we are slowly on the downward trajectory now.

It won't be slow for long, once people start realizing Sama was telling them a fairy tale, and AGI/ASI/singularity isn't "right around the corner", but (if achievable at all) at least two more technology triggers away.

We got reasonably useful tools out of it, and thanks to Zuck, mostly for free (if you are an "investor", terms and conditions apply).

[1] https://en.wikipedia.org/wiki/Gartner_hype_cycle

* The first generation were those silly AI pins last year.

* The second generation will be this opaqueAI ivePad.

* And when the third generation comes out, all phones will already have whatever makes this special.

I don't think Sama and Ive are smarter than everyone else, but even if they are, I don't see how this flies.

Re:Social media platforms: what's wrong, and what's next on Thursday May 22, 2025 @01:00PM (#44064055)
After reading all the posts, I feel this is far too strongly centered on Reddit. What about the other services? They seem to be mentioned in an aside, if at all.
Re:Social media platforms: what's wrong, and what's next on Thursday May 22, 2025 @12:44PM (#44063835)
> I feel like federated social media platforms are not going to be the answer in the end -- and although its adoption has grown in the coming years, I think it's always going to lag behind others.

Care to give an argument to substantiate that? These are pretty strong claims, "in the end" and "always" have a certain finality to them, which indicates you very strongly believe that. Why?

Cool idea. This seems US-only, unless I missed something.

Is this even possible in the EU with the GDPR and its stricter rules on medical data?

Re:OpenAI to buy AI startup from Jony Ive on Thursday May 22, 2025 @07:25AM (#44060955)
Worldcoin is and always will be opt-in.

Government solutions will be opt-out, and only in the most tedious way: Leaving the country, burning your passport and becoming a stateless person. Not recommended.

If there ever was a moat, government has it.

Thanks for the (possible) clarification.

But as long as Google isn't their _only_ customer, why would Nvidia care?

Re:OpenAI to buy AI startup from Jony Ive on Wednesday May 21, 2025 @04:33PM (#44056032)
I am a serial account-hopper, so this couldn't happen to me, so perhaps I don't understand the point at all.

But what is keeping you from making a new account and rejoining the same subreddits, except perhaps losing a hundred million magic internet points?

"With few exceptions, all animal mitochondrial genomes contain the same 37 genes: two for rRNAs, 13 for proteins and 22 for tRNAs. (...) the comparison of animal mitochondrial gene arrangements has become a very powerful means for inferring ancient evolutionary relationships, since rearrangements appear to be unique, generally rare events that are unlikely to arise independently in separate evolutionary lineages."

Source: Boore (1999] Animal mitochondrial genomes. Nucleic Acids Res. 1999 Apr 15;27(8):1767–1780

Hyperlink: https://pmc.ncbi.nlm.nih.gov/articles/PMC148383/

i.e. mtDNA changes a lot less than our "normal" cell nucleus nDNA

Have my upvote for pedantic overkill.
> Ask a person on the street who their favorite painting movements are and likely every name will be more than 100 years old

I think you overestimate the publics art appreciation. The average answer will be a blank stare.

Your post strangely sounds like Nvidia primarily makes graphic cards for consumers.

Last time I checked, they couldn't produce enough H100s/GB100s to satisfy demand from everyone and their mother running a data center. And their most recent consumer hardware offerings have been repeatedly called a "paper launch" - probably because consumer hardware isn't a priority, given the price (and profit) delta.

Re:Google AI Ultra on Tuesday May 20, 2025 @06:20PM (#44046517)
The same companies who stole... sorry.. fair-used all the worlds artworks and all text on the internet to train their models are now promising you they won't steal...sorry... fair-use your uploaded data?

In unrelated matters, I have a bridge to sell you, if you are interested.

Re:Why Does the U.S. Always Run a Trade Deficit? on Tuesday May 20, 2025 @09:11AM (#44041220)
The chain of thought goes like this:

-american labour is expensive

-because of tariffs (or geopolitics) cheap asian labour unavailable

-investment in automation research

-first Gen automation is expensive

-further research

-2nd Gen automation is price-competitive with american labour

-iterate n times

-Nth Gen automation is price-competitive with asian labour

> why doesn't it happen with the chinese supply chain

The idea is that because asian labour is so cheap, there is no real incentive to invest in automation research, because everyone knows it won't be price-competitive for long.

No idea if that is actually true, but that is the argument.

It's a fancy file format conversion utility.

Am I missing something?

It seems that PCC indeed went live with 18.1 - tho not in Europe (which is where I am located). Thanks for the heads up, I will look into this further.
It seems that PCC indeed went live with 18.1 - tho not in Europe (which is where I am located). Thanks for the heads up, I will look into this further.
Looks great. Not sure how big the market is between "need max privacy, need on-prem" and "don't care, just use what is cheap/popular" tho.

Can you talk about how this relates to / is different / is differentiated from what Apple claimed to do during their last WWDC? They called it "private cloud compute". (To be clear, after 11 months, this is still "announced", with no implementation anywhere, as far as I can see.)

Here is their blog post on Apple Security, dated June 10: https://security.apple.com/blog/private-cloud-compute/

EDIT: JUST found the tinfoil blog post on exactly this topic. https://tinfoil.sh/blog/2025-01-30-how-do-we-compare

I mean you have to be really stupid to use Grok of all models for such a task, but that doesn't mean there aren't people that stupid.
What I am wondering about is - while Musk is as unsubtle as ever, and I guess this is a system prompt instruction - is there something like that (in more subtle ways) going on in the other big models?

I don't mean big agenda-pushing things like Musk, but what keeps e.g. Meta Inc. from training Llama to be ever so slightly more friendly and sympathetic to Meta Inc, or the tech industry in general? Even an open-weights model can't be easily inspected, so this is likely to remain undetected.

> maximally truth-seeking AI, even if that truth is sometimes at odds with what is politically correct

It is certainly and undoubtedly a big coincidence that his happens to the chatbot of a white South African just when the topic is in the news again due to Trump's granting refugee status to some white South African farmers.

I would like you to read up on the railroads, and WHY private companies built them.

https://en.wikipedia.org/wiki/Railroad_land_grants_in_the_Un...

The government gave them over 700 000km2 of land as an incentive. In case that number means nothing to you: That is France. Or Texas.

Re:The Perverse Incentives of Vibe Coding on Wednesday May 14, 2025 @04:16PM (#43988719)
In section 4, the author writes "... cheaper than Claude 3.7 ($0.80 per token vs. $3)".

This is an obvious mistake, the price is per Megatoken, not per token.

Source: https://www.anthropic.com/pricing

Please don't, or at least don't without a looooooooooot of behind-closed-doors trial and error. There are few things more off-putting then an AI try-hard "i am a quirky hooman with quirky hooman personality traits".
It's a journalist coming, because you said you want to talk to the journalist, because of the bad press you had before, because you fucked up.

Seems fair to take a camera.

> Trees have nothing to do with global warming

Citation needed

>Let's not personify and demonize vines

who is personifying and demonizing vines? again, citation needed

Older »