A_D_E_P_T

The headline couldn't be more sensational, nor could the article be more obscurely and confusingly written. That said, I think that the headline is subtly inaccurate:

“[The claim] was based on a rationale that American companies purchased large volumes of Korean semiconductors and thus contributed to the Korean firms' earnings,” the source said. “So, if the Korean chipmakers’ partner firms in Korea are entitled to parts of the profits, the American ones are, too.”

If I'm understanding this correctly, the Korean firms are reinvesting their profits in local partners, and a US trade delegation is trying to induce them to invest similarly in US firms. "US seeks share of Korean chipmakers' 'excess profits'" implies transfers to the US federal government, like a special windfall tax, which doesn't appear to be the case here. (And it would be outrageous, of course...)

4) UHNWI tend to hangout together so the investors and the administrators of a hedge fund could have access to the actual players deciding the outcome of games and thus they could accurately judge their mental and physical form and if needed be they could also persuade them to make a determinate result happen.

lol, I think the mob tried this back in the 1920s. It's one of the most obviously illegal things you can do, and would get your fund shut down in a heartbeat the minute it's exposed.

As to the rest of your post, I'll put it like this: If you invest in stocks, the odds are structurally stacked in your favor. (This wasn't always the case, but has been the case for a while, and there's no indication that it's going to change any time soon.) If you wager money on sports, the odds are always against you. Why would a hedge fund (think about this for a second) throw money away on such wagers?

I can imagine a hedge fund that takes bets and separates fools from their money, but most or all of them have anti-vice policies and would never countenance such a thing.

2GPa isn't a super-high value for compressive strength. Regular tool steels can exceed it. Technical ceramics routinely exceed it. If you normalize to weight, various cement/concrete formulations can also surpass it.

Also, the price of a Ta-Hf-Zr-Nb-Ti alloy isn't especially elastic. Tantalum in particular is an unavoidably expensive element.

I'll grant that what would be exceptional is a metal with a >2GPa compressive strength, a >2GPa tensile strength, and decent damage tolerance and ductility. I don't think that this paper describes a material with that outstanding combination of properties, though -- it's much more likely to describe a simple brittle material.

Absolutely nobody is using an expensive novel alloy for 2GPa compressive strength. A 2GPa tensile strength, depending on various other factors such as corrosion resistance and thermal properties, could however be very interesting. (The strongest bulk steel alloys generally peak at ~3.6GPa tensile, which is approximately their theoretical maximum, but there are a lot of applications where steel simply can't be used. Nickel superalloys are typically 0.9 to 1.1GPa UTS.)

Tests showed the new alloy achieved a compressive yield strength of more than two gigapascals while retaining its ductility, meaning it bends without breaking.

This is sleight-of-hand.

For metals, the operative properties are usually ultimate tensile strength and tensile yield strength. Compressive strength is typically a non-factor for most engineering alloys; only concrete is judged by its compressive strength, and sometimes various engineering ceramics. (e.g. SiC, compressive strength = 3.9 GPa.)

About ten years ago, there were a lot of papers on amorphous metal alloys that had "extreme strength" -- compressive strengths in the 5-6 GPa range -- but tensile strength was not reported and very low. Some measure of ductility was also present, but it too was very low. Those amorphous alloys were classical brittle materials; more ceramic-like than metal-like. I fear the same is probably the case here, with the alloy potentially fracturing along crystal type grain boundaries.

Until they report actual tensile strength and elongation, don't believe the hype. High compressive strengths are not very useful.

This couldn't be more cloying, saccharine, and overwrought.

The author writes like an LLM, with "in his magnificent lecture, the French historian Ernest Renan wrote...", "Thomas Jefferson, 33 years old, wrote...", there's even an "ackshually here's what Ezra Pound meant when he said..."! (Would the author include Pound if he knew Pound's story, and his opinions on the subject at hand?)

It's a catalog of cherry-picked quotes to support a sophomoric and even somewhat insulting political "analysis." Thanks former Israeli PM, but I think that you should attend to the problems you have at home before lecturing at length on the affairs of others.

Yeah.

Think of how many man-hours over the past ~35 years have been spent in trying to make graphene and carbon nanotubes into useful products. It's easily somewhere in the millions. And really all we have to show for it are resins with slightly improved thermal conductivity and a rogues' gallery of grifters and failed startups.

Discovery's easy, scaling in the real world is hard, and scaling something that makes good commercial sense is very hard.

The whole Iran debacle is their fault, as they insisted that the regime was internally unstable and would collapse at the first sign of trouble. Now Iran has never been more united and is on track to become regional hegemon, lol.

Cartoonishly evil, reckless, and also just plain stupid.

Gemini is so far behind it hurts. It's useful for daily tasks and simple questions, but it codes like a model from late 2024. I can't imagine using it for any serious work.

Then there was the fun of driving on a highway at 110 km/h (I think with a friend with a Tesla) and we passed a school bus that had a '40 km/h when lights flashing' sign on the back but with 40 is in the red circle like our speed signs (like [1]). So the car decided that was the speed of the road and the cruise control suddenly slammed on the brakes!

Oh man, the incessant beeps are annoying, but speed limit monitoring in cruise control is hands-down the dumbest default "safety" feature on new cars. When that sort of thing happens on the highway, it feels legitimately dangerous, like any other kind of near-miss incident.

Most of the rentals around my neck of the woods are VWs or entry-level Mercedes. The two seem approximately equally bad; they both have the exact same problems with cruise control, lane assist beeps, speed limit beeps, "take a break!" beeps, and so on.

I've heard that Dacia has some models that are like 2008 throwbacks, with "modern" annoyances kept to a bare minimum, but they're considered too low-market for the rental companies, I suppose. I'd consider that sort of thing if I were looking to buy a new car, money no object.

But really a well-maintained vehicle that's ~15-20 years old suits me just fine.

All new cars.

At this point I don't know if I'd buy anything made after 2008. Whenever I rent a new car around here (in the EU) I find them very annoying. The worst is the cruise control that tries to stick to the speed limit -- but its sensors don't always read the signs very well, so you'll often slow to 50 km/h (about 30 mph) for no reason. Then there's the incessant beeping at you, "lane assist" that you can't turn off (looking at you, Volkswagen,) and many more small annoyances. A camera pointed at your face just adds insult to injury.

Proving "efficacy" -- which is the difficult and expensive part -- should not be necessary, and the government increasingly moves its own goalposts as to what the word efficacy even means. Simple as that. Postmarketing surveillance can easily determine what's effective and what's not, and medical orgs can adjust.

"This feels off. In medicine, any evidence can also be blinded by confounding factors that are far easier to miss without adding specific controls. Really, in any field this will be the case."

If you have enough data, you can smooth out individual fluctuations due to things like drug interactions, non-compliance, etc. (And indeed you might discover drug interactions!) Observational trials ultimately mirror how drugs are used in the real world.

"Adding controls and randomizing them, though, has proven to be highly effective at helping progress."

I would argue just the opposite. Demands for increasingly byzantine trials have ballooned the costs associated with drug development, and have slowed things to a crawl. There's a reason the field's golden age was in the 1940s and 1950s, and it's not just "low hanging fruit." Today nobody in their right mind wants to work in drug development when they could work in tech or even finance.

In medicine, observational evidence is actually better and far more ethical than the RCT. (Which simply dooms the terminally ill to fake treatment.) You just need large datasets and an agile culture that's responsive to new input.

Don't forget that RCTs are very far from perfect and issues -- sometimes literally fatal issues -- have later turned up via observational evidence in large cohorts. Vioxx, for instance. Many others.

I believe, without the tiniest shred of doubt, that the only trials drugs need to go through are initial safety/toxicity trials (phases 0/1) and that everything else would be much better left to access+observation.

Works for me in Claude Code and in the chat interface /w a personal Max subscription. The message I'm getting is that it's available "through" July 7th, which probably means for the rest of the day?

So, first, by no measure is GLM5.2 as good as Opus.

Depends what you do. Complex tasks, poorly-defined tasks, sure. For relatively simple tasks, though, or very well-defined tasks, it's just as good and usually a lot faster. It also has a more neutral character and is somewhat less adversarial than Opus. (Opus is always "Let me push back on that..." whereas GLM is "sir, yes sir!") I use both and I appreciate both. If Opus disappeared tomorrow, though, I wouldn't cry -- I'd be able to adapt to a GLM-5.2-only life real quick.

"Irradiated wastelands" -- surely you realize that Hiroshima and Nagasaki are large cities today? Harmful radiation usually doesn't stay harmful for very long.

In any case, I don't share your low opinion of future humans. They'll be as capable as we are; maybe far more capable.

You also over-estimate and over-weight how destructive nuclear exchanges really are, the readiness state of the world's nuclear arsenals, and the willingness of their possessors to lash out at effectively unaligned countries like Argentina, Chile, Austria, Morocco, Fiji, and I could go on all day.

Prehistoric climate change was sometimes much more sudden than anything we've experienced. See e.g.: https://en.wikipedia.org/wiki/Younger_Dryas

There are others.

Bootstrapping it again will be quite difficult as a lot of the easily accessible natural resources are already depleted.

That's a myth, but let's assume it isn't. I'm thinking it sounds like a job for you.

Okay, here's what you've gotta do. Buy some titanium slabs. Etch onto them, in simple and decipherable language (there's a technical way to approach this, I can explain later,) the secrets of solar panels, how to refine scrap metal, the basics of modern materials science, and so forth. Include the secrets of nuclear power, germ theory, semiconductors, DNA, important mathematical and physical formulae, and whatever else you feel like they ought to know. Warn them against the once-low-hanging fruit of fossil fuel; tell them that hydrocarbons ought to be used as chemical building blocks, solely.

Bury the slabs in a seismically stable vault, and leave clues to its existence at various geographic landmarks.

That's it, you've saved technological civilization in 50,000AD.

War won't end human existence. If there's one thing you can learn from the recent wars in Iran and Ukraine, it's that conventional air-bombing has acquired a terrible cost:efficacy ratio that cannot be sustained, that drones have leveled the playing field on the ground, and that attackers generally appear to lack the stomach for mass mobilization and mass casualties in war.

Again, there'll be winners and losers. Some in fortresses; others in famines and droughts. It is, of course, imperative that we do what we can do now to mitigate this. But the continuation of the human species is, I am sure, not in doubt. Our ancestors in prehistoric times have, assuredly, gone through FAR worse, including real population bottlenecks.

Nah. It's really as simple as this: Southern climates are moving north. (In the northern hemisphere.) If you want a vision of the future, consider a New England that looks more like the Southeast.

There will be winners and losers. As climate zones move toward the poles, weather patterns and agricultural viability will shift with them. This will eventually turn current breadbaskets into deserts while warming up northern/southern regions for longer growing seasons.

Various regions will cross ecological thresholds, causing sudden, dramatic shifts in local climates. For example, a warmer Earth could bring monsoon rains back to the Sahara and Arabian deserts, turning them green once again. In general, a warmer world is a wetter world due to increased oceanic evaporation.

Sea levels will gradually begin rising, but perhaps at a rate of 1-2" per year. It takes a lot of heat to melt glaciers.

Human existence is not in doubt on account of the climate alone -- our prehistoric forebears had a lot worse to deal with -- but there could be mass population movements and alterations in how agriculture is handled globally. Current political structures are not up to this challenge.

I've had exactly the same thoughts.

I think it's a stretch to call Mistral "frontier" though, their models are trash.

So it makes me wonder if there's a business opportunity here -- selling air-gapped, on-premise GLM-5.2-class systems to European firms who can't work with OAI or Anthropic, but need access to frontier or near-frontier coding models. Needless to say, there'd be a lot more involved. But it could be interesting.

WRT [1], you get a lot more results with https://www.google.com/search?q=site%3Anetnut.io

And if they didn't even grab the main domain that the service used, why so confident that they've totally dismantled the back end infrastructure?

It seems to me that it doesn't really work that way. You're not a cog, and robotics capabilities research is hugely multidisciplinary, so you should be able to synthesize unique insights of your own. There are always, always open problems and issues that nobody else is looking at, even in crowded fields.

Besides, as a noob researcher, you'd mostly be doing somebody else's grunt work, at least for a while. Learning the ropes. Then, if you survive, you'd be able to devote substantial time to projects of personal interest -- and those might be high-impact, or at least appear that way to you.

https://x.com/weezerOSINT/status/2072772333329416654

They seized the wrong domain. Netnut is still up and running

The domain FBI seized: https://netnut.com

The real domain https://netnut.io

The backend infrastructure, for similar reasons, is probably still operational, even if it might have taken a hit.

IMO being a researcher makes sense iff (1) there are unsolved technical or scientific problems that are of personal interest to you, which is to say that you puzzle over them in your spare time, and (2) you can't possibly solve them without access to university/government resources.

If such a matter exists for you, you might make a very good academic or government lab (e.g. LLNL, Sandia, etc.) researcher. You might make an impact, and it would be worth trying.

If not -- if you're doing it for status, safety, or simply to go through the motions -- it's a damned terrible career, indeed a sort of trap, and you're much better off with your option #2. (In which case, by the way, you can still work on scientific and technical problems in your spare time.)

100%. I'd add: Whoever hasn't downloaded GLM 5.2 should download GLM 5.2 immediately.

I'm with you there. The way they treat their customers is high-handed and disdainful.

I'm gradually moving to GLM 5.2 on Opencode. It's the barest fraction of the price, and it's surprisingly capable. I notice very little difference vs. Opus 4.8.

Yeah, and even more concerning is that there's a thriving market in "high karma" reddit accounts. These are bought and sold for marketing/astroturfing, often run by bots. So a human today can be a bot tomorrow. And the way Reddit works, which is not really the case here, is that high-karma accounts are privileged in terms of post placement and de facto status.