ecocentrik

The comparison to the poorest 3.5b is also not as useful as comparing the top 0.1% to the top 10%, which I would imagine pollute orders of magnitude more than the top 0.1%. I'm sure they're both shocking numbers compared to humans living sedentary agrarian lifestyles.

Are "principles" how this administration justifies its less than principled behavior or how you justify rewarding that behavior?

It's more a self reinforcing side-effect of a withering democracy than the cause. The rhetoric from the far right keeps implying that we're much closer to a single party state than the current election cycle suggests and they have proven themselves to be weak predictors of outcomes they don't directly control so I'm sure more shenanigans will be revealed shortly.

Three prosecutors and MIT leadership were responsible for the charges. Most US Attorney's are scummy attack dogs because their work requires it, so I would blame MIT for wanting to make an example of him instead of handling it like a disciplinary matter and a teachable moment.

AI has always seemed like a convenient panacea for the excesses of accelerationism. An overloaded term in the equation that should balance everything, but only if the right conditions are met. That framing discrepancy alone should indicate that AI is being marketed as something that it will clearly never be.

I doubt someone like Thiel would be deluded enough to think of AI as anything other than an economic accelerator and wealth amplifier.

There was plenty of network hacking going on pre-internet and people still figure out ways to hack air gaped computers. Security is always going to be a problem.

It's not minimalistic in the way you're suggesting. Their are tons of extensions listed on pi.dev that give you all the extra features you could want. You just don't need them. The harness is very capable with just the 4 primary tools. The minimalism is a claim on the harness architecture, not the number of lines of code or the capabilities of the harness.

What really doesn't make sense to me is setting uniform token targets for all engineers. I'd be more interested in setting token ranges with hard limits for certain roles that see diminishing returns from the overuse of coding agents and where maintaining institutional knowledge, architectural awareness... is more valuable than moving as fast as possible. Like everything else, these limits should be adjusted over time as tooling for those roles improves but it really feels like there are human limits that organizations should avoid exceeding.

Politics is frequently a game of semantic drift. The application of labels with unclear meaning to further dilute meaning or extend it for the purpose of engagement, legal reclassification... it's not engineering and you can't expect well defined terms from people who want power without accountability.

You don't see how those conditions amplify each other?

Most major industrialized nations have gone through something like this and it usually takes much longer than 10 years to escape and correct.

It isn't. It's about convicting generational groups, particularly the younger ones that don't know any better, that they have specific preferences.

Why should open weights correlate with cost? Cost correlates with the expense of running the model more than it does to the expense of developing the model.

Social media in it's current form is already dying to the explosion in synthetic interaction but the internet could survive and reshape itself.

It would be so simple if businesses could choose their competition or simplify disallow modes of competition they found inconvenient. I can't imagine any amount of regulation would make that possible in today's competitive landscape with the existing laws and the enormous volume of prior art.