koolba

> Evan Hubinger, Anthropic's staff lead on keeping the technology aligned with human goals and values, backed up Coxon claims in a follow-up post of his own, though he didn't quit the company.

> "Jacob is correct here — we really do earnestly believe AI could kill all humans," he said.

> Hubinger estimated the chances of that happening to be higher than ten percent within the next decade, and added that there's no plan yet on how to keep AI aligned with human goals in the superintelligence scenario.

10% chance we kill everybody is a small price to pay for Motown remixes of classic 2pac songs.

My favorite movies review was for a film in the late 90s. The name of the film escapes me, but at the time it had full page ad in the paper with quotes from the review. It said “Watching X like a beach vacation”.

I randomly stumbled upon the full review and the rest of the line was “… and you’re not invited”.

A corporation is just one or more people pooling their resources for legal and accounting purposes. Theres nothing inherently virtuous or evil about the concept.

Going back to my original example, why is it okay for a media company to use their platform and resources to promote whatever speech they want promoted, but I can’t use money I made selling hot dogs to do the same?

The only way to make the playing field fair is to allow everybody to take part. Otherwise you have a handful of media companies with outsized control of indirect messaging.

> I don't know, if we'd had an overwhelmingly Democrat-dominated government, we'd have comprehensive bans on government insider trading

We had two years under Biden with Democrat control of the House and the Senate. They passed no legislation to ban insider trading.

It probably didn’t help that the poster octogenarian for insider trading was Speaker of the House

> and we'd correct the bad ruling in Citizens United.

CU was absolutely the correct ruling as it just means I can take money from wherever I’ve earned it and spend it on publicizing my message.

Prior to that you had to buy a media company to have the same effect. It was perfectly fine for Fox News or The NY Times to focus articles and segments on what they deem newsworthy to push their narrative. But if I made my money selling widgets, I couldn’t spend it the same way.

It’s a perfectly reasonable ruling.

> Corporatism is definitely a bipartisan issue, but at least Democrats recognize it as a problem, while to Republicans it's a feature.

The same people probably complain the Senate is an antiquated body that should not exist because mob rule is the only true voice of the people.

> Data retention. Our new system of Enterprise Frontier Safeguards (EFS) gives customers complete privacy (the same as a zero data retention policy) while still being state-of-the-art at preventing adversarial use. EFS works by storing data in cloud infrastructure controlled entirely by the customer, not Anthropic. It will be made available to enterprise customers in phases, beginning later this fall. Until EFS is available, eligible customers will be able to use Fable 5.1 with zero data retention.

This is interesting. I wonder if customers will be allowed to create an auto expiry for their own data to prevent future subpoenas. That’d be a treasure trove for discovery.

The prefatory clause of the 2nd amendment’s usage of “well-regulated” means “well-equipped” (i.e., they should have lots of guns!).

It’s been repeatedly established through multiple SCOTUS interpretations that it’s just one example of why such rights shall not be infringed.

> Its like being kosher. Some people choose to not eat non kosher food. I choose not to consume AI generated materials.

This is great analogy when you consider the sheer deliciousness of pork and shellfish.

Yes it’s only bad when Trump does it.

Refusing to enforce drug laws or immigration laws is perfectly fine. Ditto for refusing to honor the 2nd amendment.

There’s trade offs for performance, access patterns, throughput, latency, concurrency, workloads, rigor, type systems, extensibility… really every characteristic you could imagine.

It’s not a solved problem because each iteration of technology doesn’t just fix the mistakes of the past, it’s an evolution to solve the problems of the present.

My complaint is not with the approval numbers. Those go back and forth.

It’s with referring to half of the country as morons. If that’s the starting point for a debate it’s not likely to be a sincere one.

Spending money on don’t have with no financial repercussions is known as the elusive “infinity money hack”. Which clearly does not exist.

Of course there’s going to be a a cost to spending beyond your means.

> It’s the people that are the problem. These policies and politics will outcast Trump because middle Anerica and The South are filled with morons.

What a sad and tired take. Trump won both the electoral and popular vote. Discounting the opinions of over half the country as morons is exactly the type of behavior that got him elected.

No deal is a bad deal. Just for an indefinite term.

> Need an end table that is lacking?

End table? That’s a server rack: https://wiki.eth0.nl/index.php/List_of_IkeaRacks

People who add sugar to coffee miss out on the sweetness of everything else by contrast.

The bitterness itself is enjoyable, but the natural sweetness of everyday foods is also magnified.

Everybody was already lazy. Now they have a means of demonstrating it.

Pandoc is awesome. My fav usage is configuring git to use it to normalize binary docs to markdown (like a .docx) so they can be diffed. Works amazing for redlining contacts.

> A good goal as a developer is a high commit message to change ratio. The more lines of code that you change, the more comments explaining why you’re changing the code. More message and less code changes per commit is generally a good idea.

No! Blanket statements like this is how you end up with 40 pages of slop AI comments in PRs that nobody reads.

Comments should be terse and meaningful. They should document surprising behavior or choices. The less comments you have, the more meaningful each one becomes because your time and eyeballs are limited as well.

Why couldn’t he have used the same laptop many years after its release?

The open source version of this has existed for decades as well.

There’s the doers that actually enjoy the craft, want to share what they’ve done, and contribute to things they genuinely find intellectually stimulating.

And for each of those true believers there’s many more fly by night wannabes that think it’s some kind of shortcut to social validation.

Turning off synchronous_commit gets you most of the way without ever worrying about data corruption if something crashes mid way through.

https://www.postgresql.org/docs/current/wal-async-commit.htm...

When I have a truly difficult prompt, I give it to the laziest model.

All of this boils down to an extra underscore in his username. It’s the type of thing I’d hope our LLM-powered future solves for the common man.

Catching typos like that would be load bearing for someone armed only with a public defender.

Be funnier if you even write “do not enter 123456”.

Damn straight. 99% of these types of plays is being willing to run in before there’s a set up. Big up to a 59 year old that is still willing to do so.

One can only hope his own knees up that well at that age.

That “-I” really needs to be lowercase.

> It’s a Herculean task, though.

I believe it’s spelled Sisyphean.

Kill a man you’re a murderer. Kill a thousand, you’re a conqueror.

And for most of those commands add a dash dash so that nothing with a dash prefix turns into options.

> will we eventually be discovering new forms of math that will in turn have some applications down the line in engineering or biomedicine etc?

If we do it probably won’t be for a long while. We’re barely using math from a couple hundred years ago for most applied usage.