I went to Best Buy to try a couple of models, just assuming I'd buy one. I was shocked to realize the "Copilot" was just "crappy skinned ChatGPT in a web app." I sort of just assumed basic agentic features would be available, like modifying PC settings. But nope. It was mind-bogglingly shockingly bad. Worse, I think it's pretty clear that "agent manager" is the correct interface for AI -- it separates "the user interface a human uses" and "the interface for interacting with an AI." For some reason, Microsoft is still shoveling out "embed a chat window in every app" as opposed to having a proper agent manager built into Windows. It's a big missed opportunity, because they could have just rolled out the red carpet for Anthropic, Google, and OpenAI to have unified OS integration built directly into Windows, but nope.

Shhh! Don't let them find out that x^2 + 1 = 0 has no solutions at all. It's our best kept secret. You have to learn, like, at least three secret handshakes and get an A in the entire calculus series before you can learn about that.

You suggested a few techniques that work in a cooler environment, but trust me, Texas sized heat will laugh at your European passive cooling strategies, and probably your improperly sized mobile AC.

In the United States, the AC build out reduced heat related deaths by 75% since 1960 [1]. At the same time, you can also have a situation where increased demand for AC creates an increase in CO2 production, which creates a feedback loop [2]. So you're not entirely off your rocker. Still, evidence from the United States establishes pretty clearly that ACs are effective for heat adaptation. That said, you say it's "for the vulnerable," but where do you confine the vulnerable for months at a time when the world heats up? You're basically saying your society, as a whole, is going to be off-limits for literally months for large portions of your own population. Proper long-term planning means you need to expand out renewables enough to intercept increased heat related AC energy demand, while considering equitable access to habitable environments for everyone. It also means you need to de-stigmatize AC use, because that's ultimately going to shame people away from AC that actually need it. At the moment, you're already seeing 25,000 in excess mortality due to the heat. That is a clear signal that "too hot" is already arriving.

Incidentally, I froze my butt off at work today, so you're right that Americans have an issue with abusing AC settings.

[1] https://www.journals.uchicago....
[2] https://www.nature.com/article...

You don't need to buy or run the AC now. But you do want to think about how climate control needs will change in 30 years. Judging from this thread, a lot of yall couldn't get an AC even if you wanted one. That's the sort of thing people need to start planning for now. But if planning for the future is too American for you, well, carry on.

Climate change is not going to stop even if the world hits the most optimistic targets. It is going to keep getting hotter, a lot hotter. In 30 years, today's peak heats will happen on 5x as many days, and the new peaks will reach 10 degrees higher. Your "a couple of weeks out of the year" is going to turn into months, and the new peaks are going to be lethal. You need to start getting your houses and culture ready now.

An insecure server is the street equivalent of parking a car with the window rolled down and valuables in sight. Someone *will* steal your crap if given a clear enough opportunity. The catch is, actually finding and exploiting bugs requires a high level of talent, time, and dedication, so it mostly suffices to make the bugs that do exist sufficiently hard to exploit. That is, at its most fundamental level, security through obscurity. The problem is, AI changes the talent and time issue.

The issue is there is a large class of bugs that were never patched out, they were just hidden because they were slightly beyond the human attention horizon. Now that AI has started to improve, that horizon is effectively machine-extended, so bugs that sat unnoticed for 8 years are suddenly surfacing. It isn't that they weren't important, it's that the bug was effectively secured through obscurity, and now that obscurity is gone. Therefore, we need to patch the bugs the AI can spot, because an attacker using an AI can spot them now. The evidence is pretty clear now that AI can find these bugs and exploit them end-to-end, so it's only a matter of time before criminals obtain models with such capabilities.

Starbucks more or less wanted to be an online order fulfillment hub. People order online, pick up, leave immediately. Ultimately, they realized this was a mistake and rolled the changes back. I have a feeling T-Mobile may end up doing the same.

Granted, I use Google Fi, so I actually *prefer* not dealing with a cell phone carrier's store. So maybe it'll work out for them.

Gambling arises from using a variable schedule of positive reinforcement based on the exchange of goods to derive a dopamine hit. Pokemon cards can be used as gambling, so can the stock market. The question is how it's being used. Someone who invests in the stock market long-term is not getting a dopamine hit from a variable schedule of positive reinforcement, so they're not gambling. Someone who plays Magic the Gathering at a gathering by buying boosters, then assembling a deck (a common play mode, both players end up with a random spread) is not seeking a dopamine hit derived from a variable schedule of positive reinforcement. But once those activities become centered around seeking the next hit, it becomes gambling. Of course there is a gradient -- someone might actually like poker, but when a major factor in the enjoyment is the risk/reward pattern, it's gambling. It becomes an addiction when it interferes with one or more major activities of daily living.

Hypothetically, someone could use a "prediction market" completely detached from the variable schedule of positive reinforcement. In practice, I highly doubt that is how it's being used, and I think deep down you know it.

You mean a tax on young and impulsive men who need guidance on making good choices? Because it's mostly young guys who are still shaking off puberty who get their credit scores ruined by stuff like this. Isn't nature beautiful? The little chick just barely waddled out of mommas nest and an alligator up and eats it. After all, the little jackass should've known about alligators!

The risk equation is:

risk = offensive_capability * jailbreakability

According to Anthropic's own doom-saying, Mythos could literally end the whole-wide world overnight with a series of catastrophic, hypercane level cyberattacks. Currently, Fable 5 and Mythos are the only Mythos-class models available publicly. Thus, even if other models have high jailbreakability, the risk level is down-regulated by having lesser offensive_capability. Again, this is according to *Anthropic.* Obviously, jail-breaking GPT 3.5 isn't going to end the world. According to Anthropic, a jailbroken Fable 5 or Mythos is the apocalypse. Since those are the ONLY apocalypse models available, jailbreakability is uniquely severe for those models in particular. This makes Anthropic's attempt to say "but GPT-5.5!" particularly moot, since no one has claimed GPT-5.5 can usher in the cybersecurity apocalypse.

Anthropic made the claims, not me. I'm just following them to their logical conclusion.

I know a doctor who fits prosthetics. He has a machine that can take scans of patient anatomy when fitting new equipment. Apparently, sometimes the 3D scan goes wrong, and he has to take it again. According to him this is especially difficult for children, because getting them to sit still during the scan can be tough. I don't know precisely what goes wrong in all cases, but some of it sounds like faulty computational geometry algorithms.

But hey, you programming people can tell the wriggling kid your project is safe from dependency hell.

By all means, avoid dependency hell. But aside lower speed: expect your solution to fail on certain edge-cases you may not realize exist, in ways you don't anticipate, when you don't want it to, and that it may be essentially impossible to identify what the problem is. It's a recipe for a legendary bug you make a workaround to "fix" that no one will ever understand during the project lifecycle.

I'm a mathie that specialized in programming adjacent things.

My main takeaway for programming people is: math is like encryption. Don't make your own. Really, even something as harmless as the quadratic formula shouldn't be coded in directly, for "reasons." Just trust me, use a pre-built library function to get what you need.

That is all.

The problem is the human brain didn't evolve (or wasn't created) to contend with an essentially infinite supply of Doritos. A lot of hyper processed foods are designed specifically to be "craveable" and hook people. Therefore, a lot of people end up locked into an unhealthy lifestyle they don't want, much like a drug addict ruins their life even though they want to stop. See, a drug addicts bears some personal responsibility for their choices. But we also routinely give long sentences for the dealers who sell drugs with full knowledge they're killing the addicts.

I think there's a deeper story here.

I ended up teaching at two colleges I previously attended. In both cases when I was granted Canvas access I came across everything from my previous enrollment, years later. I was able to find the very first messages and activities I submitted in Canvas my first week of college. Instructor hordes data it really should just delete. Consequently, I suspect Instructor paid the ransom because they know the breach exposed data of not just current students and faculty, but all past students and faculty. Any one course only exposed a small amount of information, but taken together, hackers basically got a dossier on me.

Really, if Instructor just deleted old courses and messages after a year, I might have had only a year of professional activity exposed. But as is, I'm at a seriously elevated risk of identity theft and spear fishing now. As a minor consolation, I do receive free credit monitoring and identity theft insurance as a job benefit, so I guess there's that.