doghut

We haven't had any new models from China for a while. Just a bit ago, they were shipping every other week. Here's what happened. I genuinely was (and still am) rooting for China because cheap open weight models benefit everyone and help keep the closed labs in check. The amount of capability that they were able to pack into low-parameter models that could run with limited compute was impressive, but it's now abundantly clear that China's AI labs are struggling. It's no secret that they distilled Anthropic's and OpenAI's models, but how much weight that truly carried was previously unknown. It's known now: the answer is apparently a LOT. The AI race is at its zenith, and there's essentially no other reason for the recent drastic slowdown with Chinese AI labs. Now of course, there is the chance that they're gatekeeping their most frontier models internally, potentially due to political or business-related reasons, but this is unlikely. Much of China's success comes from dominating specific sectors on pricing, which is why Chinese dropshippers dominate Amazon and Walmart. It's the reason China's the world's most successful producer, because no one else can compete with their pricing to quality ratio. Their strategy with AI is similar. Ultimately, they can't afford to not release their models, because it's the premise behind their entire business model, and from a political perspective, China also sees significantly more benefit in releasing said models versus keeping them behind closed doors. Anthropic and OpenAI have essentially fixed distillation, and clearly this is having a massive impact on Chinese labs. I'm still rooting for them and their open weight models, but without true compute, things are looking bleak.

The Department of Education under My Presidency Not all, but many teachers in the status quo are grossly incompetent and cognitively impaired. Students are often left defenseless when it comes to unjust decisions from educators that negatively impact students academically, in an unfair disciplinary regard, or otherwise, especially within incompetent K-12 public institutions that bend over backwards to defend their faculty above all else without question. Under my administration, students will have access to a line of defense powered by the US federal government, staffed by experts that can represent/defend the student and take administrative action against the educator at hand if needed. In university settings, I'll also mandate that every lecture be recorded and made accessible to students, in addition to a new piece of legislation that requires for curriculum to be fully covered within lecture (not supplemented with a textbook outside of class, not with additional dedicated overtime to learn the curriculum—everything needs to be adequately covered in lecture, or the curriculum must change accordingly).

BE the change that you want to be. SEE the change that you want to see. BUILD an AI INITIATIVE, based in MULTI-TRILLION DOLLAR AI DRIVEN SILICON VALLEY, for the betterment of the FUTURE in the STATUS QUO. INSPIRED by ARJAV MEHTA.

Don't show loyalty towards either OpenAI or Anthropic, or any other AI company for that matter. AI innovation in the status quo depends on competition. Being loyal to one specific "lab" (they call themselves that) stifles competition because an entity with a devoted following need not make additional efforts to improve (because there's no point if people follow you no matter what). It was largely due to OpenAI's challenge with Astra that Anthropic was pressured to release Opus 5.5, a model superior to Fable 5.1 in capability, intelligence, speed, and pricing/usage (so basically everything). In two days (September 29th, OpenAI DevDay 2026 in multi-trillion dollar AI driven Silicon Valley), OpenAI will give their response to Opus 5.5. They'll either succeed tremendously, or fail disastrously. That's how the AI sphere works; there's no in-between. But regardless, this is good. We WANT for OpenAI to make an effort to respond (and vice versa, when OpenAI has the lead, like it did with Astra a couple weeks back, forcing Anthropic to respond). The back and forth ping-pong between OpenAI and Anthropic is ultimately a good thing. We need to keep this competition going to help ensure the positive progression of models, general development, and to prevent the birth of a new elitist cast with exclusive access to the frontier. While many will argue that most of their user base (at least expense wise) is corporate and thus regular users are ineffective in impacting supply and demand, and hence theoretically have no impact on the companies' decision making and overall path, I'd have to disagree. API pricing isn't exactly cheap, and companies know this (they've already been cutting down on AI use because it's no longer nearly as "free" as it once was), not to mention there's the obstacle of zero data retention policies and other confidentiality issues. Regular users are incredibly good for further training of their models, collecting data for improving their products and harnesses, and so forth. They need the regular users, and that's why they're willing to subsidize the subscriptions to such a drastic extent. We're valuable to them, and in many ways, especially in terms of data, more valuable than corporate customers. Data is gold in the AI age. Show no loyalty or obligation towards either company. Fuel the competition.

Surf needs to IPO NOW. Surf has an internal self valuation of BILLIONS 👐 and BILLIONS 👐 of dollars in the status quo. Under the current private model, @zakaria is under immense strain from extreme deficit spending. It's TIME to take this to the NASDAQ. Surf needs to be NASDAQLISTED.

The downfall of K-12 education is at the fault of the educators. Having been at my university for about a year now, it's difficult to grasp the sheer contrast between a proper high-quality education (where the professors are very clearly there to serve and help the student to the best of their ability) and my previous high school (which, given my current reference point, was unacceptable on so many levels). From my four years at Silver Creek High School, a public secondary educational institution based in multi-trillion dollar AI-driven Silicon Valley, it became abundantly clear to me that 50% of the educators were fully well-intentioned, capable, and very well deserving of high praise. The other half, however, were grossly incompetent and cognitively impaired. Many were emotionally unstable, incapable of teaching (correctly), power-tripping, and severely egotistic, seeing themselves as "higher" and "smarter" than their students by some delusional and nonexistent metric. Often, the "I've been doing this for X years" (yeah and you still suck at your job), "you will not do this in MY classroom" (it's not your classroom; it's the taxpayer's), denying mistakes during lectures and shoving the blame onto the student, or frequent use of "vulgar" language (I have nothing against swearing, but frequent cursing in an allegedly professional educational environment as a faculty member showcases, at minimum, some degree of instability and poor self control) are strong indicators of an educator being in the latter 50% group. Education, at least in large, is not declining due to increasing phone use or lower attention spans. Instead, we should shift our attention to the educators themselves. Perhaps students no longer pay attention because they found that The Organic Chemistry Tutor and modern LLMs are far better learning alternatives than incompetent morons who aren't worth allocating attention towards. Perhaps education is becoming worse simply because the people teaching are idiots. We desperately need mass-layoffs in K-12 public schools.

hi