This actually works reliably with quite a few companies, especially large ones with low marginal cost services. They will often have a standard script where they will offer anyone calling to cancel a large discount for 3-12 months. People can, and many do, call back at the end of each promo period to say they're cancelling and refresh the discounted rate.
> The National Science Foundation (NSF) has put a cork in its grantmaking pipeline after BILLIONAIRE Elon Musk’s Department of Government Efficiency (DOGE) set up shop at the agency this week.
Really, "billionaire", that's what you're going with (emphasis mine)?! Isn't it more relevant that he founded and runs 2 biggest startups (as in, high velocity, high growth companies) in the world (Tesla & SpaceX) that run circles around both legacy companies and government agencies? So, yeah, if you want things to radically improve fast, of course you call someone like Elon.
It's sad that "unbiased news" basically no longer exists.
It's obvious if you see measures like web pages getting deleted because they contain the word "privilege". You know, as in "privilege escalation". I'm sure the people at NSA who wrote these pages are happy about their work being in vain.
One person blogged about `git master` and many people agreed that it may be needlessly antiquated. So private companies chose to take a few hours to change a dumb default string under no external pressure and certainly not under duress of government action.
The fact you conflate these two demonstrates either a clear lack of earnestness or common sense. Take your pick.
Did you fall for rage-bait articles about items of no consequence written to convince you that you're the victim of some mass woke conspiracy and everything you do is justified as long as you try to burn it down?
Not really, I've been opposing sexism, racism, political correctness, science denial (yes, burn it all down!) and supporting meritocracy and freedom of speech since early 2010s, way before it became big in the media.
Well you are doing a poor job of conveying it if your two biggest pain points are "git master" and "quantum supremacy".
God help us this is the way all of these conversations have gone over the last 8 years.
Group A: "You are gutting our scientific research infrastructure, a entity that has provided tremendous benefit to scientific progress worldwide and has been a source of national pride!"
Group B: "Yea well some tech companies voluntarily changed the name of their git branch from 'master' to 'main' so you all had it coming."
There's nothing I can do with that, except hunker down and hope all the horrors of this are felt by someone else.
I didn't know Tesla ran circles around legacy car companies. Well other than P/E ratio. I was under the impression that they made fewer and worse cars that are currently not very popular.
What would the unbiased version of this report look like? "...after SUPERMAN Elon Musk's...", "...after NATIONAL HERO Elon Musk's..."? In what way would you like journalists to show their deference to this great man?
Billionaire is incredibly relevant. It highlights the inherent conflict of interest in the moneyed class demolishing the state that is meant to protect the rest of us from them.
I'm not sure how it was done legally, but when the Supreme Court ruled that Biden administration's student loan forgiveness was unconstitutional, Biden (or rather, his team) found another way to forgive loans.
I expect Trump's administration to similarly find legal workarounds.
Is it any different, in principle, to what congressmen (and women) have been doing for decades? Insider trading, corruption, etc. has all been normalized.
You’re forgetting that the previous government was losing at least 2 wars (in Ukraine and against Houties) and destroying the United States - trying to jail political opponents, subverting elections, destroying the country’s borders, erasing meritocracy, instituting censorship and ignoring Supreme Court rulings.
Those topics weren't in the scope of the original discussion and I'm not really interested in litigating partisan entertainment propaganda based around taking shreds of truth (at best) and blowing them out of proportion.
The large scale facts are that under the previous administration we had working relationships with our allies, mostly functional executive agencies (aka law enforcement), and the US (ie USD) was seen as a source of stability. Meanwhile the current administration's actions are indistinguishable from a foreign power doing its best to destroy our country - we are now isolated from our allies (and even seen as hostile!), the ideal of rule of law has been replaced by brazenly corrupt rule by law, and we're staring down dedollarization.
> individual tokens are routed to different experts
that was AFAIK (not an expert! lol) the traditional approach
but judging by the chart on LLaMa4 blog post, now they're interleaving MoE models and dense Attention layers; so I guess this means that even a single token could be routed through different experts at every single MoE layer!
They clearly mention, take into account and extrapolate this; LLM have first scaled via data, now it's test time compute, but recent developments (R1) clearly show this is not exhausted yet (i.e. RL on synthetically (in-silico) generated CoT) which implies scaling with compute. The authors then outline further potential (research) developments that could continue this dynamic, literally things that have already been discovered just not yet incorporated into edge models.
Real-world data confirms their thesis - there have been a lot of sceptics about AI scaling, somewhat justified ("whoom" a.k.a. fast take-off hasn't happened - yet) but their fundamental thesis has been wrong - "real-world data has been exhausted, next algorithmic breakthroughs will be hard and unpredictable". The reality is, while data has been exhausted, incremental research efforts have resulted in better and better models (o1, r1, o3, and now Gemini 2.5 which is a huge jump! [1]). This is similar to how Moore's Law works - it's not given that CPUs get better exponentially, it still requires effort, maybe with diminishing returns, but nevertheless the law works...
If we ever get to models be able to usefully contribute to research, either on the implementation side, or on research ideas side (which they CANNOT yet, at least Gemini 2.5 Pro (public SOTA), unless my prompting is REALLY bad), it's about to get super-exponential.
Edit: then once you get to actual general intelligence (let alone super-intelligence) the real-world impact will quickly follow.
Well based on what I'm reading, the OP's intent is that, not all (hence 'fully') validation, if not most of, can be done in-silico. I think we all agree that and that's the major bottleneck making agents useful - you have to have human-in-the-loop to closely guardrail the whole process.
Of course you can get a lot of mileage via synthetically generated CoT but does that lead to LLM speed up developing LLM is a big IF.
No, the entire point of this article is that when you get to self-improving AI, it will become generally intelligent, then you can use that to solve robotics, medicine etc. (like a generally-intelligent baby can (eventually) solve how to move boxes, assemble cars, do experiments in labs etc. - nothing special about a human baby, it's just generally intelligent).
Not only does the article claim that when we get to self-improving ai it becomes generally intelligent, it's assuming that AI is pretty close right now:
> OpenBrain focuses on AIs that can speed up AI research. They want to win the twin arms races against China (whose leading company we’ll call “DeepCent”)16 and their US competitors. The more of their research and development (R&D) cycle they can automate, the faster they can go. So when OpenBrain finishes training Agent-1, a new model under internal development, it’s good at many things but great at helping with AI research.
> It’s good at this due to a combination of explicit focus to prioritize these skills, their own extensive codebases they can draw on as particularly relevant and high-quality training data, and coding being an easy domain for procedural feedback.
> OpenBrain continues to deploy the iteratively improving Agent-1 internally for AI R&D. Overall, they are making algorithmic progress 50% faster than they would without AI assistants—and more importantly, faster than their competitors.
> what do we mean by 50% faster algorithmic progress? We mean that OpenBrain makes as much AI research progress in 1 week with AI as they would in 1.5 weeks without AI usage.
To me, claiming today's AI IS capable of such thing is too hand-wavy. And I think that's the crux of the article.
Yeah I think the math+code reasoning models, like o1 and r1, are doing what can be done with just pure compute without real world validation. But the real world is complex, we can't simulate it. Why do we make particle accelerators, fusion reactor prototypes, space telescopes, year long vaccine trials? It's because we need to validate ideas in the real world that cannot be done theoretically or computationally.
I don't get your irony. Yes EU companies are smaller, do the fines are proportionally smaller, but they are held up to the same standard and fined as well. It's not some hidden scheme to extract money from the US.
The limits are conveniently set so that the law doesn't apply to most EU companies. Only 4/25 included companies are EU (and 3/4 of those are porn, Booking.com isn't).
Edit: it's definitely worse if you go deeper into the rabbit hole. Sister legislature, Digital Markets Act:
Booking.com insisted on the fact that it is one of the only European companies that is a global success and that as they are not the most dominant actor in this sector, they should not be disincentivized while competing with bigger companies.
So yeah, "please only punish non-EU companies" definitely sounds like a trade barrier.
For what it's worth, I bought a $2000 robotic hand from US a year ago, and paid about €400 in VAT and €32 (so like 2%) in import duties / tariffs.
Obviously VAT isn't a "trade barrier", if anything it's a "consumption barrier" and it's the same for every business that EU citizens give money to (i.e. if I bought a robot hand that cost $2000 to make from an EU company, I'd likewise be paying €400 VAT on top of that).
I should (pretend to) cancel more often!
reply