Survivor bias

I’m unconvinced that morality is a simple Boolean. He might have ethical standards that are different that others though.

Yes. I'm not sure whether Amodei's motivation for attempting to ban open-weight models was for safety or to strengthen Anthropic's position though.

Depends on the objective of Apple. It’s hard to imagine they’re after a quick payout. They may wish to keep this in the news cycle as long as possible. I could see them both harming open ai and sending a message to employees thinking of leaving that if they even consider breaking their confidentiality agreements, it will absolutely ruin their careers.

doubt it - in the uber/waymo thing the guy got pardoned by trump after lobbying from thiel/luckey, probably a similar outcome for this guy since altman has trump's ear rn.

I mean, Trump turns on people in an instant depending on his interests.. he's not a reliable friend. He is, also, very thin-skinned and he (probably?) knows his base is extremely against datacenters. (It's hard to ignore that political reality now -- it's like the nations' most politically unifying thing!) Dario was in the admin's good graces, until very suddenly he wasn't.

Good point. If these are the ethical standards they go by, who’s to say they’re abiding by any standards to keep my data private.

You get about 80-90% of the results for daily tasks like: getting summaries or explanations of complex material. Writing software tools for data analysis. Getting recipes for a given set of ingredients in the fridge.

> 80-90% of the results

why would I ever want 80% of the results though?

The reason you never hear that is because waste and fraud are very low in reality. That’s why this made the news. It’s uncommon.

We should be using the government to help people and when we do, it often does a good job.

Examples: Roads, libraries, fire departments, schools, safety regulations…

Our public transportation infrastructure literally cost 10x per mile than France or Hong Kong. That's not waste to you? For what California has spent/is spending on high-speed rail from nowhere to nowhere, China blanketed their country?

Notice you also left out police. How's our spending working there?

California high speed rail has been a mess. Which is likely in part because politics dictating routes has raised the costs and timelines substantially. On a smaller scale SF was building a new subway in 2013 that had been on the drawing board for years. I remember thinking maybe I'd ride it to work one day. Opened in 2023 or 2024, after I had moved offices twice and then went to work from home. It's not a terrible line but because it had to go to the center of our Chinatown instead of 2 blocks over, it took quite a bit longer and became the deepest subway line in the city. Several other bits of stupidity too in that project but a big piece of the delayed timeline was that tunneling in SF is hard.

Plenty of other transit projects exist that have made real differences. Personally I'm a fan of the simple improvements: revised bus routes with dedicated bus lanes and improved stop & shelters, added bike lanes, etc. those sorts of projects are relatively cheap investments and while no single one is a silver bullet they add up. On a bigger scale - Caltrain's electrification was a big win. Both kinds of projects are easier than building whole new tracks or digging new tunnels. Extend and improve the existing systems. Most cities have something to start from.

> Personally I'm a fan of the simple improvements

Stockholm Syndrome. Because our government is so paralyzed and ineffective, we can't even dream big anymore.

I mean I'd be a fan of a roadmap of simple improvements that contribute to a big vision. Shipping incrementally just seems to work better when possible.

Note Atherton's delays cost Caltrain electrification 3 years and $400m IIRC

https://peninsulaforeveryone.org/blog/atherton-spent-145k-to...

I’d note that China does not need democratic approval for any of its projects and thus is substantially more efficient.

We don't either. It's a Democratic Republic. Our electeds could do their job.

And what lesson should be drawn from that?

That it's easy to build things when you oppress and enslave people.

We seem to get the worst of both.

I mean building infrastructure is literally done by private groups the government hires not the government itself. Infantalizing the government as unable to do anything is the biggest myth of the 20th century. When the government did build the roads for the interstate high system they were much cheaper. They cut through any possible regulations and just built them. If government is slow its because we have regulated government to be slow because people would get tremendous power and than plow over large African American communities with roads and rail roads. Its very well documented the speed and brutality that government can move with.

All this to say the government can be fast and efficient mostly because we have an almost infinite number of examples in the US of it doing it.

> because we have an almost infinite number of examples in the US of it doing it.

... before the 1970s. Nothing this millennium, or even the last quarter of the 20th century. It's been, at minimum, 50 years since we've cost-effectively built anything at scale.

The police in America is amongst the best in the world. They do a good job keeping people in line projecting authority and terror.

> The reason you never hear that is because waste and fraud are very low in reality. That’s why this made the news. It’s uncommon.

Fraud might be relatively rare, but we hear about waste constantly and it isn't newsworthy because a high level of waste is expected from a government operation. If it wasn't then there would be good results by putting things under government control, rather than the more realistic outcome of a politically controlled market oscillating erratically between gluts and shortages depending on who won a popularity contest this year.

To be newsworthy political waste generally has to be competitive with wartime levels of destruction. Things like a 2 billion dollars going down the tubes with basically nothing to show for it. Which is similar to the cost/benefit to the US of a day of wild flailing in Iran, US assaults typically run at around a billion dollars per day. If it is less damaging than that then there isn't much point bringing it up because there are bigger fish to fry.

And another, our recidivism rates are much higher than comparable countries. Is that how we do a "good job" helping people?

The cloud services wrote the contract and the UI for their console. They then encourage young developers to try out their tools and encourage a market environment where those skills are needed to secure employment. Some kid goes and tries to build their first web app, they follow instructions and tutorials but miss that a single default selection on a menu three nested layers down is going to cost $2,000 per month. This isn’t disclosed on the page. Sure, it can be determined by reading several different documents, but the provider chose to not show estimates for costs in the setup.

How is that the kid’s fault?

Nuke the data. It’s gone forever if you didn’t back it up elsewhere. This should be a meaningful risk mitigation that I can employ to avoid having a catastrophic financial disaster.

This isn’t a limit I’m setting at some percentage above expected costs, it’s: “I don’t want to take out a HELOC if something goes wrong”

Unfortunately, a lot of people keep their backups in the same cloud account as their primary data. Thinking that multiple copies and multiple availability zones are sufficient.

For these users, the article’s €54k bill would be replaced with their business data getting wiped out.

I would love to have a “if the bill for this hobby project becomes a threat to my ability to pay my mortgage, nuke it.” If I cared about the data enough. I’d have backed it up.