andyjohnson0

No. I think it'll go all the way and we'll end-up with something approximating agi.

Reason is that the tech is hooked-up to the dying husk of late capitalism like an iv line, and nation state military competition will force it to work. I don't see it being restrained much, unfortunately.

So dotcoms are the wrong analogy. Computational intelligence is like the internet - its an everything technology. People (like me) won't like it, but I think it will be in everything. Peculiar alien minds are just gonna spill out of this thing.

They can't fix it because the thing that they need to fix it is the thing that doesn't work. /s

But seriously: while I don't use Claude, this issue of perceived unreliability seems to be approaching the point of existential risk for Anthropic. Whats the theory about why they're struggling? Compute capacity? Load? Lack of focus on SRE?

Put it another way: is their downtime due to something fundamental about serving inference, or just bad engineering choices? Given their resources, it seems astonishing.

I once interviewed with a company that organised their people into four "tribes". The company culture was pretty odd, and I fortunately managed to bomb the in-person interview. Now I know where they got the idea from.

Point taken. But I'm not an academic and this is just hn - and I think the comment was well made.

Edit to add: The critique in the linked blog post refers to weak studies relied-on in one chapter of one general-readership book by Kahneman. I'm not aware of anyone claiming that he is generally unreliable as a scientist.

Three things:

1. Many people intuitively assume that clumping/clustering of events implies non-randomness, and that random processes are smooth and low-variance. The opposite is true [1].

2. A consequence of 1. is that people often over-estimate their understanding of the likelihood of events and the degree to which they are conditional/dependent.

3. There was an intriguing comment on this site a few days ago [2], referencing Daniel Kahneman's work on System 1 and System 2 thinking. From memory it said that reality is a lot less explicable than we tend to think - and that a lot of what we casually think we know about the everyday world is just our brains filling in the gaps using quick and cheap System 1.

As to why people are clutching at science-fictional interpretations: perhaps they're looking for some excitement or novelty? That would be very human.

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

[2] Unfortunately I cant find the comment. I wish I'd favourited it.

One of my colleagues has a mechanical keyboard - possibly a Filco - that they use in preference to their crappy corporate-issued one, and i have come to loathe the machinegun sound. I get that they feel good to use, and the haptic and audible feedback combo is particularly effective. But for the love of $deity I just want the noise to stop.

Firefox has solid and capable cross-platform sync

I think what your parent commenter means is that, if the application is warlike or nefarious, them the money will be found. If, on the other hand, it is humanitarian, then every penny will be counted.

> Travelers going between cities/countries carry items for people who need to send stuff.

Carrying othet people's parcels across national boundaries is a really, really bad idea for the person involved. Drugs, money, weapons, explosives, endangered species, etc.

If you plan to run a business that facilitates this then you are exposing yourself to potentially very severe legal liabilities.

It seems to me like magic links are just off-loading the auth problem to email.

If magic links become increasingly common then email† account access increasing becomes a single point of vulnerability/failure. And email providers obviously can't use magic links for auth. I dont know what the solution is.

† or IM services

Interesting article. I learned some things.

How hard would it be for Mr Github to add rss/atom feeds, I wonder?

Thank you - I hadn't read that before. Its a much richer, and also darker, work than The Last Question.

Also it was written in 1980,.almost three decades after The Last Question. I wonder if part of the difference (to me) is in the evolution of the author's writing practice, or development of themes in SF over that time?

> strip few least significant bits

I'm unconvinced. Doesnt this just replace the need to choose a suitable epsilon with the need to choose the right number of bits to strip? With the latter affording much fewer choices for degree of "roughness" than does the former.

The triteness was more in the ending than the overall exposition. Humans create computer, computer creates universe->humans.

> I suspect you've read a lot of works derived from Asimov

You're probably right, although the transitive chain of derivation is necessarily long. Clarke - probably not derivative. Blish and Cherryh (some), Stapeton, Lem, Heinlein (the juveniles, as a kid), Baxter, Banks, Gibson, Ken MacLeod, Charles Stross, Peter Watts... I dunno.

I did grind through the Robot books as a child, and the Foundation books that he wrote. But just because they're foundational (no pun intended) doesn't stop them feeling stuffy and dated now.

(And as an aside, it strikes me now that Clarke's The Nine Billion Names of God is kind of the anti-particle to The Last Question.)

I'm going to make myself unpopular here, but I've never understood the perennial gushing about this story on hn.

The writing is okay, but the ending is kind of trite (especially given the author's humanist beliefs. And there's much too much exposition.

Convince me I'm wrong.

Stewart Brand

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

The OLSA is far from perfect, particularly in the definitions it uses, but its focus is social media platforms and the like, not simple blogs.

OFCOM has a q&a-based tool [1] to advise on whether the OLSA applies to a site. I'm not a lawyer, but its pretty clear that a non-adult-themed personal blog where people can post only textual comments on content that they dont control, is not going to be subject to the act.

[1] https://www.ofcom.org.uk/online-safety/illegal-and-harmful-c...

Do one thing at a time, and do it well.

Human brains didn't evolve to be proicient at cognition-heavy multi-tasking. You can habituate yourself to it to some extent, but there's a good chance that your exhaustion is your body asking you not to treat yourself like a machine.

From here in the Uk the site just says:

"Unavailable Due to the UK Online Safety Act [...] Now might be a good time to call your representatives."

So I fired-up a vpn, and it appears to be a personal blog. About ai risks.

The geo-block is kind of a shame, as the writing is good and there appears to be nothing about the site that makes it subject to the OLSA.

I completely agree. But this argument would be incomprehensible to the class of people who own Google and similar corporations.

> Seems to be a very regular occurrence starting around this time of day (14:30 UTC)...

8.30am on the US west coast

This might be better posted on LinkedIn. HN isn't your blog.

Account's comment history didn't look particularly AI generated. And, as an organic human who uses em dashes myself, I kind of hope we can get past this simplistic take that they are a signifier of ai content.

Besides that, I thought the comment had something useful to say — whether ai-generated or not.

Not the OP, but I interpreted "religious" as something like a strong, unyielding preference. As in "I wear my helmet religiously when cycling."

That would be the Sam Altman who left YC back in 2019? Or some other, more recent, Sam Altman?

Not sure why this was apparently flagged to death. Vouched.

The next US election is 2½ years away. This about what is happening now.

Quote:

"Effective immediately, the United States Navy, the Finest in the World, will begin the process of BLOCKADING any and all Ships trying to enter, or leave, the Strait of Hormuz," he said.

I'm looking at doing some field recording. Currently reading around the subject and thinking about starter equipment.

Report Card:

USA, #1 world power, nukes etc, failed to win. Iranian regime wasn't destroyed, and the HE uranium is still there. Two carrier battle groups weren't enough to re-open the straits. Burned $45bn and achieved nothing much.

Iran, fourth-rate regional power, failed to lose. Regime still in power, and still has its HE uranium. Oil infrastructure intact. Military knocked-about a bit, but still capable of controlling the straits and bringing in the $$$. They get to keep brutalising their frightened population.

Israel, premium regional power, suddenly all on its own running a hot, two-front war with no easy exit, because Trump cut and ran. Netanyahu is probably toast in the upcoming elections, which means hes going to jail for all the corruption stuff.

China, mega military power, picked the right side and comes out of it clean. Looks like a model, dependable global citizen. Access to the oil without paying the $2m/boat straits fee. Didn't blow $45bn on nothing.

Russia, third-rate power on the way down, made some roubles on the oil price thing, but still has all the same problems as before. Meatgrinder war. Economic and demographic collapse still imminent. Global pariah, except for Orban and Erdoğan and the Belarus guy. Putin's getting pretty old.

EU+UK basically sat it out. Sensible. Much economic damage though.

Thousands of dead/wounded/traumatised adults and children -- a running sore that shames us all. Plus all the victims of the Gaza and Lebanon wars. Deep, deep anger for multiple generations. Revenge, etc.

----

Officially its a ceasefire, but everyone knows the US won't restart after this. Its six months to the mid-terms, and the genius orange guy in Washington has other things on his mind. And Epstein is still hanging around like a bad smell. What to try next?