hungryhobbit

This assumes any research gains last. In practice it seems like none of the frontier companies can keep much of their "secret sauce" secret for very long.

If company X has a better model than company Y, do you use it? Well, that depends: will Y take two years to catch up? If so yeah, you buy X.

But if Y is just as good a few months later, you purchase based on other factors, because you know you can live without the brilliant research X is doing ... since you'll have it soon enough.

The problem here is we're not talking ... I don't know ... hamburgers. If I like McDonald's better than Burger King, it's not going to stop me from buying a Burger King burger (if I'm hungry and there's no McDonald's around). One burger may be a bit better, but I can reasonably substitute one for the other.

Coding AIs aren't burgers. Even if a local LLM is as good as Opus 4.5 (when you sink several thousand dollars into hardware ... and I'm skeptical even then) ... the actual difference between the models is huge.

Let's say I use the $200/month Claude (I don't, I use a cheaper plan, but stay with me). If I switch to a Opus 4.5 equivalent local LLM, and it makes even a single production issue that 5.5 wouldn't, that mistake will cost thousands of dollars.

That "cheaper" model isn't really cheaper.

This, so much! Newsom has demonstrated, with his actions, that his words (and political actions, like this) cannot be trusted.

This is the man that told the entire state, millions of people, to shelter in place ... and then went to one of the fanciest restaurants in the entire state (French Laundry) to solicit campaign donations during lockdown.

Newsom will say or do anything to become president, but he should never become president precisely because he will say or do anything, with no moral compass guiding him.

Who is this source? It seems like Ukrainian propaganda to this outsider.

(Don't get me wrong, I'm rooting for Ukraine! But reading either Ukraine's or Russia's take on the war means reading PR ... and not reading anything I'll actually learn from.)

It's cool, but fundamentally has the same issue as not using it: there's too much data, and no way to interpret/act on it.

What the whole thing needs is an AI-based summary of the judges at the top "This judge rules more favorably to ____ and ____, and rules against ____ and _____ more often than most judges".

I understand a human who parks in a disabled human's parking space and then gets mad at him for being late to a meeting is an asshole who doesn't deserve to be emulated.

Society has nothing to do with this, as pretty every member of society would agree the above is atrocious behavior. What you're saying is that the entire human system is bad about rewarding people at the top appropriately ... which is absolutely true, but its not about people wanting it that way; EVERYONE (except the 0.0001% at the top) wants it another way ... we just can't figure out how to get there.

>There's an important distinction here. Anthropic isn't actually spending more cash on AI than salaries. It's valuing its internal AI consumption at retail prices.

If virtually any other company in the world said "we have internal AI servers and we let our employees use them" ... the company would measure that as they would any other internal expenditure: by how much it costs them. You blow $50k a year on servers running your LLMs? It cost you $50k.

Only a company like Anthropic, desperate to make news and help their IPO, would measure by the cost they charge customers. It's as ridiculous as Microsoft counting every employee's use of MS Word ... as if it had paid the full consumer price of each copy.

The irony of Sam Altman's company pretending to care about what's "fair" is so very rich.

He literally made billions exploiting the intellectual work of millions of his fellow humans, yet has the audacity to whine about how the system isn't fair.

Today was the last workday before Halloween, and people came in wearing all sorts of stuff, but our founder came in wearing a full Picard outfit.

I was like "nice Halloween costume" and he responded "Halloween? I just wore this for Patrick Stewart." I almost teared up.

Plus, in long-running series like these writers (implicitly and explicitly) write for the actor. It's a bit chicken and egg-y, but you absolutely cannot (honestly) pretend that another actor would have the same lines written for them.

I mean, in season one sure, but by the time you get to "there are four lights" Stewart and Picard really are inseparable.

IDK about the "being an asshole to your face" part (my story on that, above, did come from early Apple).

But he was still famous for working his teams an insane amount of hours throughout his career. It was an open "joke" that any new project he released would be completely unsupportable for the first week ... because he drove the entire team that built it so hard, the entire team crashed for a week after.

That is not how any healthy/reasonable adult runs any team, tech or otherwise.

Less "forgetting" and more "not as brilliantly insane as Jobs" ... but Jobs was NEVER someone that ANYONE should ever emulate.

Fun real life story: I worked with an engineer once who had some sort of disability that required him to walk around with crutches attached to his arms. Before he worked with me, he worked at Apple, and was high up.

One day he had a meeting with Jobs, and he was late, because a non-handicapped car had parked in the handicapped spot, and as a disabled person it took him a long time to get from a normal parking spot. He said as much when he arrived in the meeting.

Now here's the "Jobs is burning in hell right now" part. Jobs looks the disabled guy directly in the eye and, without any shame whatsoever, says "oh, I parked there ... I believe the US government overreacted after WW2 and built too many disabled parking spaces, so I park in them."

He says this ... to the sweaty disabled guy who just had to "race" across the parking lot to get to the meeting! Fuck Jobs and fuck emulating him.

Its the fundamental problem of our age: everyone is building pieces of a system, and almost no one is actually looking at how those pieces fit together into a system.

This is not about (just) the healthcare system ... it's the core problem with:

- our legal system (millions of laws and rulings and zero effort spent reconciling them into a good system that treats everyone equally)

- our economic system (we reward people who create true wealth in many different ways ... but those rewards are wildly disproportionate to how much true wealth they add)

- our tax system (again, millions of laws, and yet the people who should be paying the most pay the least because of how they fit together)

- our electrical system (although, due to its centralization, this one is actually catching up much faster than the rest)

- insert any other system here

I imagine it would look like a bunch of North Korean guys (probably not starving like their countrymen, but probably not looking great either) sitting in something like a cross between a NK state prison and an American office.

So yeah, really not exciting ... just sad.

They are trying!

The Japan thing just reflects the book's age.

It was published in 1992, but it was likely written in large part during the 80's ... and back then the world thought Japan was the future because of the popularity and success of Japanese tech corporations.

Their failures a few years later would dispell that myth ... but for awhile it really seemed plausible, and you'll see a similar fixation on Japan in lots of sci-fi from the 80's.

Honestly, IMHO it's complete and utter crap compared to the AMAZING, INCREDIBLE work Stephenson did with everything that came before.

Snow Crash? A milestone in modern sci-fi. Diamond Age? A truly incredible work of art! Even lesser stuff like The Big U or In the Beginning Was the Command Line were really, really good.

Then came Cryptonomicon, and it was good too ... but he was already deteriorating into loving the sound of his own voice way too much by then, so you had to wade through a bit to get to the good stuff.

And then there was Anathem ... where he made up half a language and required you to learn it to read the book (and it was like ten zillion pages long). Good writing is NOT about just throwing whatever brilliant ideas you have onto the page ... it's about cutting it down to the good stuff. Stephenson just stopped doing that by the time he got to Anathem.

To be fair, lots of people were willing to wade through all that fake language BS and Stephenson writing 100 words when 10 would do, and they all report loving the book ... but as a complete and utter Stephenson fanatic, even I abandonded him at Anathem.

They paywalled their site. That means they don't want sites like Hacker News sending people to them.

"Cram your square pegs into this very round whole we built!"

The link is broken (it should be https://github.com/LiberNT/LiberNT ... and really it should be the project's website, since it has one).

It's evil to call Lake Ontario ... which has had that name for over three centuries (almost four) ... the same name it has always been called?

The article asks "Why so cautious?" Well ... maybe because half our country would rather believe what one guy with a long, long history of dishonesty says over ... you know ... reality?

Now, to be clear I still 100% fault Google for giving in to them (and for renaming established geographic landmarks like Lake Ontario). The "do no evil" company has become the "how evil can we get away with" company, and that's not even a recent change.

But why? That part seems pretty clear.

TLDR; Alphabet hit negative free cash flow for the first time ever since going public.

Translation: even Google ... a company that basically has a license to print money (aka Google Search) ... is losing a ton of money on AI.

This seems to assume 3d printed homes will never develop technology that makes them superior to conventional homes ... can you see the future?

Seems like maybe they just didn't watch long enough.

Just because Anthropic isn't dumb enough to enshitify their lead model in the first month doesn't mean they aren't lowering performance over time.

You're really just realizing:

>that "crime is legal" for his circle...

now?

I don't follow how "we can see tiny stuff" = "it is possible to build closed system"?

Seeing tiny stuff doesn't magically give you enough info for Claude to reason with the system predictably (the way it can with source code).

Does it matter? It's like saying "we fireproofed 1 of 200 apartments in a building" ... well great, but you still can't sell that building until you fireproof the other 199.

Same issue here: no one in America can even buy these things (practically-speaking), let alone drive them legally in US streets until China starts building them to conform to US safety standards.

Until that happens, we're just seeing nifty demos of impressively cheap cars ... that are also (subtly) very dangerous.

It's almost like there are different-sized Discords out there, and some people (e.g. you) use only the small ones, while other people are different, and some of them use large Discords.

To be fair, it's not that simple, and having read the article you should know as much instead of painting this false argument.

What ELSE did they wipe?

>the phone was rife with “numerous sexually explicit messages and images” sent to the victim as well as “significant personally identifying information about her” is not clear error. The record reveals that these conversations occurred over multiple apps on the phone and that Stathas also received explicit photos of the minor. The court did not clearly err by crediting the government’s assertion that “no procedure short of digitally wiping” the phone would eliminate all criminal or harmful material.

So a convicted child molester used multiple phone apps to contact his victim, and instead of going into every app on the phone to see if it had child sex abuse evidence the police just wiped his phone. The court said (more or less) "that's not unreasonable: cops have better things to do than hand wipe criminals' phones of criminal evidence".

Now is there a larger question here? Sure ... but lets not pretend police clearing evidence of child sex abuse from a phone (which the convicted child sex abuser claimed contained evidence ... for a lawsuit that very likely would never happen) is as simple as "police deleting evidence of their wrongdoing".