No. RSI doesn't mean a model modifies its own running code/weights. It trains another copy of them as a baseline for the next round of the experiment.

They are. It's just not happening on the versions we have access to. Look at Sakana. RSI just involves defining an experiment and parameters, and varying the values of the parameters, trying out a training run, and record the results. If it works well, use the results as part of the base model for the next iteration. All of that can be automated. I have no idea why people are saying they can't be. If people iterate on them and change one variable at a time, and use that as a means to improvement, an AI can be tasked to do the same. And they can also be trained to design the experiments themselves.

You left out "trade secrets". That would be plausible if the software were sufficiently innovative.

That's "tool using", not "tool making", but I wouldn't be surprised if dogs could to simple tool making. A cow was recently documented making a backscratcher.

If rolling a styrofoam ball is using a tool, adapting a towel to a flyswatter is making a tool. I'll admit that I've never met someone who was that limited. Everybody used to be able to whittle something, if not something fancy, but I think that knives are now generally not used that way. I don't even know if rubber bands are still used to launch folded paper (which is multiple instances of tool making).

From outside, an IR imager should be able to see the heat from a leak. But who knows; one would imagine if it were that easy they'd have done it 7 years ago, when this crap Russian module's chronic leak issues began. My inner cynic thinks this is probably political; competent people aren't being permitted to deal with it.

They'll deal with it when some hatch or docking port blows out and sucks a couple people into space wearing NASA tee shirts. I'm rather certain about that much.

I've never met a human who couldn't make a tool. Not a complex tool, but just a tool. Complex tools are a lot more difficult, but anyone who can adapt a towel to a flyswatter can make a tool.

... big thing. I don't think anybody has anything against any vehicle, tractor or other, or anything at all stuffed to the brim with useful electronics. (emphasis on useful) The problem is when that technology is proprietary, disfunctional on purpose and designed to be extortive. That farmers are sick and tired of that I can see clearly.

One big part of the problem also is that farmers are locked into their business harder than other people, more prone to corporate extractive and extortive business pratictices and they are likely not the type to have the free time to deal with these practices in other ways.

Setting up a non-profit and/or publicly shared business to offer hardware designs that counter these problems are a likely candidate for some use- and helpful businesses. I expect this to be the next big area where the FOSS concept catches on.

To a constitutional literalist, much of what the federal government does is illegal. Unfortunately, the Constitution, if literally interpreted, would not work in a large society with fast transportation and fast communication. It was written for a country that was 90% rural, and where it could take weeks to travel to Washington, DC. It was also written for a country where most of the decisions affecting citizens were LOCAL. Town or county level.

And, yes, the government has clearly drastically altered to interpretations put on many sections, and ignored others. This was necessary because the Constitution is too difficult to amend. (Perhaps they would have done it anyway, but that's a separate argument.) E,g, there is no valid basis in the Constitution for any law either enabling or regulating a corporation. The founders generally didn't trust corporations, and to the extent that they were permitted, thought that they should be regulated a the state level or lower. Some of them thought that corporate charters should be for limited periods of time...but just imagine trying to build a transcontinental railroad without standardized legal corporate laws. (It's been done in other areas, but it sure wasn't easy.)

While true, that's like saying dirt can't exist without gravity.

Capitalism existed long before markets existed, and markets existed before people did. Fish cleaners staking out a site for their business in the ocean is a market. An amoeba storing resources for later use is capitalism in action. (I.e. it's getting stuff now for later use, accumulating capital.)

Most of the things that people attribute to capitalism are only the property of one "dialect" of capitalism. And corporate capitalism is itself a cluster of dialects, that exist under specific legal constraints, which vary with time and place.

So, yes, markets cannot exist without capitalism, as markets are about exchanging stuff, and capitalism is, basically, "the way one handles stuff". I suppose one could rephrase that as "the belief in the way one handles stuff", which would eliminate amoebas, etc., because they don't practice belief, but the way it's commonly used doesn't seem to imply belief.

While money != capital, capitalism isn't basically about improving anything. It's about using the capital you have (i.e. stuff that's under your control, which includes money) to increase the amount of capital you have. This often improves things for at least some subset of the people, but that's a side effect, and if it's missing, what is described is still capitalism.

It's only that now, roughly 25 years late, even the dimest of dimwitts in the political sphere have noticed that proprietary software is shitty by design and expensive and thus plan to move to FOSS rather than continue spending trillions of Euros on software that experts have downloaded for free and in better quality from the intarwebs for decades now. One should never say never I guess.

It's only by coincidence that that software (mostly) happens to come out of the US. Which is totally beside the point of why FOSS is gaining traction anyway. FOSS from the US will certainly be part of that transition too.

It may even be true. But the article I read didn't say how many qbits they had on the chip, or what the error rate was.

There are already commercial quantum computers from DWave https://www.dwavequantum.com/ . They're rather limited, but they exist. But I don't expect them to be general purpose turing complete by 2029 with very many qbits. Possibly by 2035. And I don't expect them to be personal computers until they stop needing to be cooled with liquid gasses (i.e. supercooled). But I wouldn't bet that this will never be possible.

I expect that there are groups in the government that have reasonable expertise in that area. But I see no evidence either of what they are (probably some folks in DoD and NSA might have relevant expertise) or reason to believe that they would be tasked with the review.