Why jump to regulation when just simple law enforcement would suffice. All of these OpenAI "rogue agent" events have been illegal, but no DA is enforcing them.

I think enough of these improperly constrained agent events have occurred that we can safely say this is misconduct of a level necessitating serious and concerted regulation of AI labs. We can't wait until serious harm is done like the disruption of medical or social services.

The much more interesting recent news IMO is that Google is not allowing (non-Samsung) OEMs to sell devices with GrapheneOS:

https://news.ycombinator.com/item?id=49946698

See the last paragraph.

For example, non-Samsung Android OEMs aren't allowed to directly sell devices with GrapheneOS and Google will only permit it within a quota. It can and is being worked around and there will be devices sold with GrapheneOS as the stock OS without Google restricting how many can be sold.

My guess is that the workaround is that Motorola sells them with Google-certified Android. A third-party (non-OEM) buys them in bulk and preinstalls GrapheneOS.

But this is really end-90s/begin-00s Microsoft levels of anti-competitiveness. I'm surprised that (particularly non-US) regulators are not investigating them yet.

Reminds me of this: https://www.youtube.com/shorts/vHfbUIQeW_A

You move to USA from a 3rd world country, take money from the Americans, Arabs and the Chinese, hire bunch of Israeli, Pakistani, Canadians, Romanians and Russian engineers and and you become an American success story.

You found a robotics company in Germany and you instantly fail the purity test, you are American company if you took too much US money, you are actually Polish if you hired too many Polish engineers etc.

It's just weird attitude.

All my data is out there, one way or another, and a dedicated cybercriminal - or worse, a government entity - can obtain or exfiltrate it without issues. I know it, they know it, everyone knows it.

The only thing I can change now is my reaction to this fact, and although the idea of off grid autarky is tempting, I am not there yet. I just don't want to stop living - flying abroad, going to doctors - I just accept that privacy in the current state of human condition is impossible, and move on with my life.

I don't think, for a vast majority of cases, these companies I'm forced to interact with can be trusted with my data and it's having a real world negative impact. Even with the best intentions the information is somehow valuable to steal and I'm baffled how it's not secure.

There should be some consequences for companies asking for things like SSN/National Insurance numbers on job adverts or retaining drivers licence photos after test driving a car, they just don't need the data anymore.

However, people need to get it in their heads that they're not chatting with their secret BFF, they're chatting with Big Tech. Before LLMs, Big Tech had no way to scrutinize the bulk of what was going on within their services, so you could have a secret hate diary in Google Docs. Now, everything you say or write can be automatically screened for red flags on a planetary scale, and probably will be because that's what the regulators and "concerned citizens" will demand. In a couple of years, you'll be biting your tongue a lot more often in private chats.

> The master said: "Now you are truly enlightened."

The average Nebraska farm is 989 acers[1] as of 2022 and uses roughly ~1,200 acre-feet of water (or roughly 390 million gallons) per year[2]. So a single average Nebraska farm uses roughly 30 times the water the Lincoln data center uses per year. There are 44,479 farms/ranches covering 44 million acres, which accounts for over 90% of the state's total land area.

Until I looked this up, I didn't appreciate the scale of water usage in farming (even in a state like Nebraska that gets a large amount of rainfall compared to Central or Southern California). I hope that the current data center environmental focus continues to shed light on our industrial water and electrical usage.

[1]http://farmflavor.com/nebraska/nebraska-crops-livestock/top-... [2]https://cropwatch.unl.edu/2024/2023-irrigation-and-water-man...

New AI data centers are obviously compute-oriented instead. That means they consume a lot more power, like hundreds of megawatts. For perspective, on average a U.S. household uses only 1.2KW. Because of this, cooling needs are significantly increased, and while water usage is often overblown, noise volume is a major problem. At (edit) 10m, it often exceeds the 90dB threshold for hearing loss.

Then there's the fact that millions of dollars of GPUs are inside these things, so security means lots of bright lights at all times of day.

These are just problems with the physical presence of a data center, not to mention the electric bill hikes, land value loss, environmental effects, etc. And there's no obvious positives to one going up. No new jobs. No tax money because it's often elided with incentives from bribing local officials.

With all the negatives and seemingly no benefits, ask yourself who wants these data centers, and why? If the only way to improve AI is using more and more compute like this, what is the end-game?

The current data center craze just feels like the same thing but on a bigger scale.

The platform APIs were terrible. React wasn't "more fun"... it just made it possible to do things with the platform that were extremely difficult and cumbersome to get working reliably with platform APIs alone.

To me, web components were an incredible idea poorly implemented. Most of the minimal adoption happened on top of frameworks like Lit that wrapped WCs to try to make the dev experience tolerable.

In urban planning they have a concept called "desire paths" where if you don't put sidewalks and pathways in the right places, people invent their own. I feel like the web community has spent a lot of time and effort patching the platform.

Now credit where credit is due: it has very much improved. And modern standards means it really is time to re-evaluate where and when you need these patches. But don't write off all that annoying and painful effort people put in to trying to making the platform deliver it's promised potential just to "building it yourself is more fun".

Invariably, the following happened:

- I was given grunt work as a I was the new/junior guy

- Did this for about 1.5 weeks

- After the above, someone (usually the owner's wife) would say something like: "Instead of Alex doing grunt work, why doesn't he helps us "

And that ranged from "order us a new computer" to "fix the North American Controls lumber sorter machine".

My point is: Being able to solve complex problems is a surprisingly transferable skill. It's so transferable that even in the trades, if you can solve those problems, people will identify it and rapidly move you to the most complex problems they have.

Probably for companies and opportunistic individuals with a high risk tolerance. As for me, just give me hard budget caps. It shouldn't be a choice between supporting one of the approaches, when you could let each client pick what they want.

Want an alert? Configure one.

Want a hard limit? Configure one.

(just make the configuration easy and visible)

When you’re more worried about this sort of stuff vs generating profit then the history books strongly indicate that’s usually not a good omen of what’s coming next.

So the focus completely shifts from spending 90% of the effort on the functionality of feature A to spending 99% of the effort figuring out how to safely implement even a lightweight feature A.

I've worked physical labor too, though never in a factory. More like lugging baggage and cleaning toilets. Still, people got hurt doing it, especially over a lifetime of labor.

It comes down to rarity of skill. Not to be to crass about it, but when your primary value is that you're willing to put yourself in physical harm, the market for your offering has a lot more supply than demand.

For example, I spend a lot of time in hotels when I was young. Just finding people who were reliable and trustworthy and legal enough to drive the airport van without wrecking it were like 1 in 10 of the people who were potentially eligible to do it. If you were able to do rudimentary math and not steal from the till, you were one of the "lucky" few who got to work at the front desk. If you showed up regularly, well, the world was your oyster.

On the note of

It was, unfortunately, a nightmare. There were tons of tickets and even threats of lawsuits from customers whose service got cut off hard at the worst possible time due to organic growth/going viral/big event/nobody knew about the limit/etc. Not only did they lose all the leads and revenue they would have gotten from that bump, but they also pissed off their own existing users who suddenly couldn't use the service either.

Generally speaking, it's much better to use alerts instead of hard limit. Even in the worst case (hackers pwn your credentials and mine Bitcoin or whatever) the rest of your business is unaffected and you can negotiate with the billing department at comparative leisure.

This is all assuming you have humans operating the service. If you're letting AI agents yolo infra in prod, you have a whole series of new problems.

Watching him try (and fail) to build a composite airplane says so much about the allure of modern techniques (I won't spoil the end, but old-school gets the last laugh). I've been less than enthusiastic about modern composites ever since seeing this show.

Bob's attempt (and short temper) reveals a good deal of hubris on his part.

The documentary is a fascinating failure and a masterclass in attempting to salvage a show when the whole premise for the show completely collapses.

[1] https://youtu.be/FSNzF-KSGiw

- If you allow to trolley to proceed, there's a 50% chance it will run over every human on the planet

- But if you flip the switch, it takes the long way around, possibly bankrupting the trolley company. And you have a legal obligation to the shareholders to prevent that from happening at all costs.

I obtained a BS in Aerospace Engineer in 1998. After job jumping several times I realized I didn't care for it. I was basically a "paper" engineer, which was my fault as I never pursued jobs that did "real" engineering.

Anyway, around 2003/2004 I went to night school for EMT. The purpose was to be a dive master on dive boats (I was/am a dive master - I don't dive anymore). It was fun.

My engineering job sucked so I jumped into paramedic school right after EMT. I really enjoyed it. About a month from graduating PM school, county fire department came in to class and said they would hire paramedic only but we would have to cross train to firefighter.

In Feb 2006 I was 35 yo with a 9 month old baby and decided if I didn't do it then I never would so I applied and got hired then department put me through fire school.

It was the best career decision of my life. It didn't take long to see a bunch of people get hurt from the job, so I figured I better have a back up plan since I also have chronic health issues. From 2010 to 2015 I did an online Masters degree in CS. After completing it I started working part time remote jobs as a developer.

Then COVID hit in 2020 and they pulled me out of the field due to my health issues. I turned 50 and was getting tired after 14 years. I started applying to full time jobs and landed one.

In 2022 I switched jobs again for a better org and I am still here. I just switched positions from Java dev to cloud engineer in May.

I absolutely enjoyed 85+% of FF/pm but it is a young person's game. It gave me confidence I probably never would have found. Again, it was literally the best career move I ever made and would recommend anyone to do it.

No doubt Timur contributed heavily to this given Valves Steamdeck (which uses a very similar but slower GPU).

Given the current hardware prices it's pretty awesome to see someone squeezing maximum performance out of old hardware!