The 4 GWh per year makes sense. The factory will produce 4 GWh of batteries every year. So if each battery is (pulling a number out of my butt) 10 KWh, the factory will build four million of them per year, or about 11,000 per day (assuming 365 days per year operation).

As for the translation to households... I can't make anything sensible out of that number. I'll note that you (inadvertently, I'm sure) multiplied the number of claimed households by a million, but if we drop that down to the claimed 4M households, that means each household gets 1 kWh of the capacity, which is enough to run the average house for about 45 minutes of average consumption... so that's not "overnight" storage for solar generation, but it's more than you need for grid stabilization. Dunno. My best guess is that they were going for how many houses you could run for a month on that much storage and just screwed up the arithmetic by three orders of magnitude (one SI scaling factor).

The $71M figure is to build a factory that manufactures sodium-ion batteries, 4 GWh of them per year.

Nah. It's just like every previous innovation-driven boom (railroads, oil, telecoms...) the growth and buildout phase generates incredible new wealth, highly concentrated (Vanderbilt, Stanford, Rockefeller, Carnegie, etc.), . Then there are several decades of adaptation and adoption, during which the benefits accrue to the population broadly. This doesn't decrease the wealth of those who got crazy rich during the boom, but their wealth stops growing rapidly and they just become the new "old money". Then the next innovation creates a new crop of winners and the cycle repeats.

We're really just completing the adoption and adaptation phase of the personal computer revolution, and maybe a third of the way through the adoption and adaptation of the Internet revolution (the dot-com boom).

This claim ignores the fact that each new boom creates an entirely new set of winners, nearly all of whom actually came from humble-to-moderate beginnings. The wealth isn't staying in one "ossified" caste, indeed it's the exact opposite of ossification. Of course, as I said above, the wealth of those winners doesn't decline during the adaptation phase, but the the wealth generated by the newly-realized productivity gains does accrue broadly -- and in fact those "winners" never actually had as much wealth as they appeared to, because their paper wealth was primarily composed of the market's projection of their slice of their company's slice of the broad future gains.

In the end, the tide does raise all boats. The US has been getting steadily wealthier -- at every quintile -- for decades. There have been some dips at each recession, but the trend line is clear. Look at the real median household income and the net worth of by the bottom 50%. Granted that inequality has risen... that's to be expected during and for the first decade or two after an innovation-driven economic boom, and we've had multiple, in quick succession. But the actual numbers show that the tide is lifting all boats -- though the boats of the big winners have rocket engines attached.

There is one very large caveat that we're dealing with right now, which is housing prices. Due to a combination of factors -- mostly dominated by increases in square footage per person and quality of amenities that we've become accustomed to accepting as the baseline, but also including excessive building restrictions and a minor real estate bubble caused by the pandemic and low interest rates, then compounded by rising interest rates needed to combat the post-pandemic inflation -- right now that increased income and wealth isn't enough for many to buy the homes they'd like. Wages continue rising and housing prices in much of the country are actually falling a little, so over the next few years that will likely equalize out, but right now it's painful for many. But that pain is generating a pessimistic view of the state of the economy and individual incomes that simply doesn't match reality.

There definitely is a risk right now that this could all go badly if the new winners are allowed to engage in extensive regulatory capture. That's how market-proof monopolies are built, by capturing and directing government power. If we can avoid that, though, and just let the market continue to work, the general wealth increase will continue.

Well, unless the AI singularity hits. It's impossible to know what things will look like on the other side of that.

Oh, you're a conspiracy theorist. Okay, have a nice day, I'm out.

What you say is true... but basically all of the other president have actually been pretty smart people. There are some who clearly aren't as smart as others, but I can't think of any who were as dumb as Trump.

Selection, not coincidence. He's the right kind of grifter, with the right kind of personality. He's really not very smart, though. If he were actually smart he wouldn't shoot himself in the foot with such regularity.

Probably LRO vs. Danuri

Yes, Donald Trump is stupid, and yes he knows exactly how to play to his base. These things are only incompatible if you think it requires intelligence and strategy to play to his base. It doesn't. He just does what he does and it happens to resonate with them, most of the time at least.

Meh. "Gambling" is such an amorphous term when it comes to betting on events that aren't carefully constructed to be the result of pure chance.

This is really no different from many other areas in which you can have conflicts of interest. For example, sports! And the answer isn't to ban betting, it's to ban betting by people who have a conflict of interest. This is well-trod ground.

I'm not particularly a fan of Kalshi or Polymarket and don't really care if they exist or not. My point is just that this isn't really any different from anything else.

The biggest clue that he hasn't read the books -- or maybe just doesn't remember them well -- is that all of the books and stories were built around demonstrating ways in which the rules failed. The 0th rule is only one example, and far from the clearest.

We don't know the exact answer yet, but IMO, you're two-halves right.

The key difference is, humans aren't "trained" primarily via a clean library of tagged data (which in itself is a learned skill). We experience the world as a more-or-less continual stream of noisy data from our several senses. We self-train how to interpret (and primarily, filter) sensory data; and more importantly, how to derive new training data from it.

To be clear, I'm not saying we have so many neurons to support input filtering (though that's admittedly at least partially true). I'm saying we virtually can't run out of new training data. Every second we're alive, it just keeps coming in.

Except instead of pursuing policies that would help open models at the expense of Big AI, anti-AI people pursue just the opposite - policies that can cripple open source AI but have no realistic ability to meaningfully hurt big AI. It's so short sighted.

Actually, no, what they usually do is much more insidious: fingerprinting rather than watermarking. The fingerprint isn't actually included in the audio, it's included in a database. If they want to tell if the track was generated, they just try to match the fingerprint in the database. Can't filter it out of the track like you can with a watermark.

The two mainstream options (both about half a year old now, we're due for something better) are Ace Step-1.5 and Stable Audio 3 Medium.

Ace Step has full vocals with its music. See here for examples. The downside in my opinion (judge for yourself) is that it has Suno's flaw of sounding too "clean", "mainstream" and "uncreative", except even moreso (Udio was always much better than Suno at this, albeit "less well behaved" - but Udio is out of the game now).

Stable Audio 3 doesn't do vocals, though it has other useful features, like inpainting (great for fixing glitches in recorded tracks for example). In my view, it sounds a lot better. It can also be used to create sound effects. License is a bit more restrictive, though, if you actually care about that, but generally won't affect the average user.

Both are trained on fully licensed data and both are open-weights. So even if the devs released a version that included some sort of watermarking in the future (or someone developed a fingerprinting algo for the existing versions), you could just finetune them to break that.

TL/DR, if you want open source and want to just churn out a full track: choose Ace Step 1.5.
If you want open source and want to do instrumental tracks or to supplement manual work (such as your own vocals) or do audio editing: chose Stable Audio 3.