Good point. Jev and Muse have been copied after 1 month of release. So happening now.

The top US labs, if their model progress is sufficient to permit its customers to “steal software”, can instead of releasing new models use them to steal software for themselves.

The tech bros have insane GPU lease agreements. More than they can reasonably afford/use. When “amateurs” can recreate photoshop, 1000x “free” (would be wasted) compute at a better photoshop they control is something they can do too, and beat OSS and adobe. There are capable decompiler to readable/organized source code agents now too. At one point, the LLM labs can stop releasing new models, and just steal everything themselves.

The moat is shrinking overall for open models, but for models that fit in under 128gb ram, passing opus 4.6 and pushing up against 4.8 is an impressive achievement.

There are 1-2gb vram models that should fit in anything. Game doesn’t appear to be texture heavy, and targetting a 6gb gpu, would cover anything less than 10-12 years old.

qwen flash is great. Privacy/sovereignty and protection from others stealing your data is very valuable and equivalence to frontier 3-4 months ago is very impressive. Your point is correct, and larger models have the capacity to solve problems with more general knowledge baked in, where smaller models simply can’t find critical information from search, or develop a strategy as well if they are clueless, and just repeatedly randomly guessing.

At the same time, they are doing extreme price gouging with little improvement between generations since h200. Huawei actually has better $/performance ratio and accelerating faster. Others have better inference solution, with AMD doing better on $/training performance too. The fact that Nvidia is the one propping up the AI bubble, sinking their cashflow into dead ends, makes it a bad stock too. Peak NVidia BS got exposed in an Australian IPO this week. It was too absurd to get anyone to buy and got cancelled. Heavily backed by NVidia.

SpacemiT K3

less performance at comparable cost to an amd 780m mini pc. lower power, but lower performance/watt even. Some future potential though. It is more of a competitor to raspberry pi, but where you need 10gb ethernet and processing beef: firewall/router, video decoding, some AI assistance/routing/gateway

The only possible achievable competition goal for US in AI is in building skynet. Boiling the oceans to obtain pure fascist oligarchist supremacy over Americans and pro Israel opinion in the world, far more than military targetting or weapons production supremacy advantages.

There is zero chance long term of having the power infrastructure to support AI/datacenters in the US, and zero chance of extending AI into manufacturing dominance. China’s manufacturing supply chain, and abundance instead of tariff scarcity, makes it impossible for them to lose their edge to anyone unfriendly to China. Models that are less than 2 months, and closing, behind, and massive south east asia datacenter development that provides 90% of local frontier US model traffic consumed in China, makes US Oligarchism’s national security pretense a farce.

US is collapsing late stage nazi colonial economy, and geopolitical AI is not even the last bastion to put all eggs in a basket for, it is a direct tool to pillage the US/colonies for oligarchy during accelerated collapse.

Hard to understand author’s point, but it is supposedly this:

When someone is sick, lonely, or frightened, do we want a real person in the room, or will a bot do?13 Do we want our children taught and shaped mostly by people, or are we content to hand that to machines? When a court weighs whom to believe, or an employer whom to hire, do we want a person behind the call, or can a machine sort it?14 Do we want the systems that increasingly run our lives steered by a few private companies, or held more widely accountable? And do we want a shared world we can make sense of together, or one so flooded with synthetic stuff that no one can tell what is real?15 None of these can be answered by asking whether AI produces quality or slop. They are about the communal lives we want, and when we reach for “AI slop,” we’re sloppily pushing each other on exactly these questions without ever confronting them out loud.

When we complain about an example of AI slop, we’re implicitly opining on all of those questions against over reliance on AI. Author is complaining that there are more dangerous/annoying aspects of AI overtaking decisions rather than content production.

It seems innapropriate/unfocused to bring up discussion of those topics, with zero opinion/argument by author, framed as a distraction from not stating an opinion on slop.

transparent in its methodology

To qualify as transparent means disclosing “based on x days in this date range”. That is not disclosed.

SpacemiT K3

what is this hardware performance and price?

Annualized Revenue rate is a made up number, that is a separate accounting fraud, for being a purely selective cherry picked time frame that gets multiplied.

The reason they are changing their made up number from $70B to $50B is because they used to include their partners’ revenue (extra $20b annualized) as their own.

The bubble part is the extremely high losses, even before hiding training expenses as assets that become worthless the next month after a new version release. Also monumental promises that they will require massive new fundraising for, that fewer people understand remain possibly sustainable.

The $20B difference is based on an accounting fraud. $50B vs $70B is like if GM would count the total of what buyers pay for cars ($70B) instead of their real revenue ($50B) of what their dealers pay them. Monumental fraudster BS came out overnight to resume the bubble frenzy though.

hint: is the baby roasting in oven video accompanied by a hasbara/description of “this is what hamas/palestinians do to Israeli babies”?

toxic sludge is good for you

This is perhaps obscure, but a reference to a famous PR book from the 1970s. Datacenter harm is mostly about electricity scarcity pricing for rest of US, noise and sometimes water use, and financial loss. There can be pollution/global warming associated concerns, but its not a major health care issue.

stop treating oligarchs like they don’t have something to gain by supporting each other.

This is indeed the core issue in America. Israel first rulership is part of the support circle, and much more so, since 2024.

That $1B budget is unlikely to be TV advertisements for datacenters. It’s likely to be bribes that get all of the pre-election candidate opposition to projects go quiet after election.

White house meeting ended up with a “morally binding” agreement to be more generous to communities/states considering data centers.

Apparently, its cheaper to spend $1B on “toxic sludge is good for you” lies than it is to offer community benefits.

After a human reviewer examined the statements, they were reported to police.

The guardrails are working as intended. ONLY persecute the public with them. Government skynet, and bond villains will have access to non guardrailed models, or any “flagged for review” markers will be ignored to bomb more elementary schools. A sheriff’s office is government power related. School shootings is just a thoughts and prayers issue, if you’re confused about the line between this and other incidents/attitudes we are being spied on for.

bookies put up their own money on the other side. adjust odds to try to even up the sides of a bet. Prediction markets are like a casino rake on a poker game. The casino is not gambling its own money.