We need an OpenOpenAI to open source OpenAI who should actually be called ClosedAI, since there's nothing open about them other than their banks to take all your money.
I canceled my OpenAI subscription last night, as did many many others. There were some threads in reddit with everyone chiming in they all just canceled too. imo OpenAI is done, and will go through massive cuts and probably acquired by the end of the year for a very tiny fraction of its current value.
You want to bet?
The panic around deepseek is getting completely disconnected from reality.
Don’t get me wrong what DS did is great, but anyone thinking this reshape the fundamental trend of scaling laws and make compute irrelevant is dead wrong.
I’m sure OpenAI doesn’t really enjoy the PR right now, but guess what OpenAI/Google/Meta/Anthropic can do if you give them a recipe for 11x more efficient training ? They can scale it to their 100k GPUs clusters and still blow everything.
This will be textbook Jevons paradox.
Compute is still king and OpenAI has worked on their training platform longer than anyone.
Of course as soon as the next best model is released, we can train on its output and catch up at a fraction of the cost, and thus the infinite bunny hopping will continue.
> The panic around deepseek is getting completely disconnected from reality.
This entire hype cycle has long been completely disconnected from reality. I've watched a lot of hype waves, and I've never seen one that oscillates so wildly.
I think you're right that OpenAI isn't as hurt by DeepSeek as the mass panic would lead one to believe, but it's also true that DeepSeek exposes how blown out of proportion the initial hype waves were and how inflated the valuations are for this tech.
Meta has been demonstrating for a while that models are a commodity, not a product you can build a business on. DeepSeek proves that conclusively. OpenAI isn't finished, but they need to continue down the path they've already started and give up the idea that "getting to AGI" is a business model that doesn't require them to think about product.
In a sense it doesn't, in that if DeepSeek can do this, making OpenAI-type capabilities available for Llama-type infrastructure costs, then if you apply OpenAI scale infrastructure again to a much more efficient training/evaluation system, everything multiplies back up. I think that's where they'll have to head: using their infrastructure moat (such as it is) to apply these efficiency learnings to allow much more capable models at the top end. Yes, they can't sleep-walk into it, but I don't think that was ever the game.
> The panic around deepseek is getting completely disconnected from reality.
Couldn’t agree more! Nobody here read the manual. The last paragraph of DeepSeek’s R1 paper:
> Software Engineering Tasks: Due to the long evaluation times, which impact the efficiency of the RL process, large-scale RL has not been applied extensively in software engineering tasks. As a result, DeepSeek-R1 has not demonstrated a huge improvement over DeepSeek-V3 on software engineering benchmarks. Future versions will address this by implementing rejection sampling on software engineering data or incorporating asynchronous evaluations during the RL process to improve efficiency.
Just based on my evaluations so far, R1 is not even an improvement on V3 in terms of real world coding problems because it gets stuck in stupid reasoning loops like whether “write C++ code to …” means it can use a C library or has to find a C++ wrapper which doesn’t exist.
OpenAI issue might be that it is extremely inefficient with money (high salaries, high compute costs, high expenses, etc..). This is fine when you have an absolute monopoly as investors will throw money your way (open ai is burning cash) but once an alternative is clear, you can no longer do that.
OpenAI doesn't have an advantage in compute more than Google, Microsoft or someone with a few billions of $$.
oh wow. I have been using kagi premium for months, and never noticed, that their AI assistant now has all the good AIs too. I was using kagi exclusively for search, and perplexity for ai stuff. I guess I can cut down on my subscriptions too. Thanks for your hint. (Also I noticed that kagi has a pwa for their ai assistent, which is also cool)
Computing is not king, DeepSeek just demonstrated otherwise. And yes, OpenAI will have to reinvent itself to copy DS, but this means they'll have to throw away a lot of their investment in existing tech. They might recover but it is not a minor hiccup as you suggest.
I just don't see how this is true. OpenAI has a massive cash & hardware pile -- they'll adapt and learn from what DeepSeek has done and be in a position to build and train 10x-50x-100x (or however) faster and better. They are getting a wake-up call for sure but I don't think much is going to be thrown away.
In my experience with deepseek and o1, openai's big talk about (and investment into) hallucination avoidance might save their hides here. Deepseek may be smarter, and understand complex problems better, but it also seems to make mistakes more often. (It's as if it's comprehension is better, but it's worse at memorization/recall.)
Need an LLM to one-shot some complex network scripting? as of last night, o1 is still where its at.
My experience gels with yours. Given the same code sample, DeepSeek has better, more creative suggestions about how to improve it, but it can't implement them without breaking the code. o1, generally, can implement DeepSeek's suggestions successfully. I think chaining them together might have quite interesting results.
That's ok if all you want to know is which model should I use today, but a test like that is totally dependent on training data, and there is no reason to expect that either DeepSeek-V3 (the base model for R1) or the additional training data for R1 is that same as what OpenAI used for O1 and whatever base model it was built on.
The benchmark comparisons are perhaps, for now, the best way to compare reasoning prowess of R1 vs O1, since it seems pretty certain they both trained for those cases.
I think the real significance of R1 isn't the released model/weights itself, but more the paper detailing (sans training data) how to replicate it, and how effective "distillation" (i.e. generate synthetic reasoning data for SFT) can be to enhance reasoning even without using RL.
The big deal here isn't that R1 makes any other models obsolete in terms of performance, but how cheap it is $2 vs $60 per million output tokens compared to O1 (which it matches in benchmark performance).
O1 vs R1 performance on specific non-benchmark problems is also not that relevant until people have replicated R1 and/or tried fine-tuning it with additional data. What would be interesting to see is whether (given the different usage of RL) there is any difference in how well R1 vs O1 generalize to reasoning capability over domains they were not specifically trained for. I'd expect that neither do that well, but not knowing details of what they were trained on makes it hard to test.
1. You can get all the models by buying Kagi subscription (excluding o1). Includes DeepSeek models. You can also feed the assistant with search data that you can filter.
2. If you have GitHub Copilot, you get o1 chat also there.
I haven't seen much value with OpenAI subscription for ages.
I have Kagi Ultimate and it is nice for this. But a cheaper suggestion would be to use OpenRouter and then use these models via Fireworks or TogetherAI. It also integrates into much more applications. AFAIK Kagi doesn't document a user facing API for the assistant feature.
Sure. I meant moreso that this would be cheaper than Kagi while providing the same selection of models.
As for deepseek, I couldn't even sign up because my email domain is not on their whitelist. To just try it out for now I don't mind the increased cost.
I disagree, I don't really need "conversational chat responses", I need multimodal
ChatGPT is the king of the multimodal experience still. Anthropic is a distant second, only because it lets you upload images from the clipboard and responds to them, but it can't do anything else like generate images - sometimes it will do a flowchat which is kind of cool, GPT won't do that - but will it speak to you, have tones, listen to you? no.
And in the open source side, this area has been stagnant for like 18 months. There is no cohesive multimodal experience yet. Just a couple vision models with chat capabilities and pretty pathetic GUIs to support them. You have to still do everything yourself there.
There is a huge utility for me, and many others that dont know it yet, if we could just load a couple models at once that work together seamlessly in a single seamless GUI like how ChatGPT works.
if the law explicitly says bytedance and there is no way for bytedance to avoid it then its a bill of attainder and unconstitutional. presumably, they have worded the law in a way that avoids this for example by letting the president remove bytedance for being in violation if he considers them no longer in violation.
No, he can't. Congress would have to revoke it. But it has bipartison support. So its just more of the same charade BS that he rants on about. Its all nonsense from him. It will be worse this time around bc he is not all there (even moreso than 2016). The next 4 yrs are going to be quite comical. He can't even control his bowels and he has to wear diapers to stop leaking.
I'm no fan of trump, but the law explicitly states that the president can exempt a platform.
> The Act exempts a foreign adversary controlled applica-
tion from the prohibitions if the application undergoes a
“qualified divestiture.” §2(c)(1). A “qualified divestiture” is
one that the President determines will result in the appli-
cation “no longer being controlled by a foreign adversary.”
§2(g)(6)(A).
> The President must further determine that the
divestiture “precludes the establishment or maintenance of
any operational relationship between the United States operations of the [application] and any formerly affiliated entities that are controlled by a foreign adversary, including
any cooperation with respect to the operation of a content
recommendation algorithm or an agreement with respect to
data sharing.”
The content recommendation algorithm is TikTok. It is developed in China by Chinese engineers. There is no lawful way for TikTok to operate under this law, and the SC has completely failed by not considering this. It's a really lazy judgement.
Probably the outcome Congress was hoping for is that it gets sold to a US buyer who would operate TikTok with the technology under license, and everyone would just pretend that now it's operated by a US interest despite that being impossible. Like sure, they would be running the servers, but the code would largely still be written in China!
Edit: Actually it would be kind of worse, because thinking about it TikTok has a lot of engineers outside China now, so how would it even work? Would they fork the code and then that would be it? It's such a crazy proposition.
Real time content recommendation algorithm can be rebuilt from scratch relatively quickly (weeks). At the beginning it won't be as effective as current TikTok algorithm so iterations will be required but frankly treating algo like something that can only be done by Chinese engineers is silly.
When TikTok developed recommendations it was novel and on the frontier but now how it's done is much better understood and with GPUs availability can be implemented by any good ML team. Similar to Facebook, Instagram, Youtube and other, the secret sauce is content and users, not algorithm.
Concentrating power in the hands of the few is certainly a good way to get an oligarchy, which is what the "checks and balances" system of the US government is supposed to prevent. It's strange to see so many people wanting the president to have more authority and power, but I guess it's a response to Congress's reputation of being dysfunctional and refusing to compromise.
As of 2024, a representative strong-executive democracy with a large authoritarian leaning and an unhealthy obsession with oligarch-worship.
In 2028, who knows. The current president told his supporters that in four years, they won't need to vote anymore, whatever the fuck that is supposed to mean.
Simple answer. A chinese owned company has no such rights or protections. Free speech does not apply. The law also does not censor content (so no free speech violation anyway). The law simply bans the distribution of the app on marketplaces stores for reasons stated (national security). Big difference.
> Simple answer. A chinese owned company has no such rights or protections. Free speech does not apply.
The Constitution does not place limits on which people are protected by it (you don't have to be a citizen for it to apply as the founders were looking to limit the powers of their government not their citizens). And with the expansion of those protections to corporations through Citizens United, I'd be surprised if a court found that `company + foreign != person + foreign` when they've decided `company == person`. (Well not surprised by this Court.)
> The law also does not censor content (so no free speech violation anyway). The law simply bans the distribution of the app on marketplaces stores for reasons stated (national security). Big difference.
The rest of your comment still stands right in my eyes. National Security has often been used as a means to bypass many things enshrined by the Constitution.
Eh? Unless otherwise specified, corporations satisfy the definition of a person across all federal laws per 1 USC §1, which reads: "the words “person” and “whoever” include corporations, companies, associations, firms, partnerships, societies, and joint stock companies, as well as individuals"
That 1 USC §1 is not a typo: this copy appears in the first section of the first title of US code, on disambiguating common terms used in law.
Totally beside the point. Verbatim from Citizens United:
> The Court has thus rejected the argument that political speech of corporations or other associations should be treated differently under the First Amendment simply because such associations are not “natural persons.”
That is not nearly the same thing as saying that they are people. Just that when it comes to this particular right, the way it is applied is not functionally different. That’s like saying that because corporations pay taxes they are also people.
The point you and others try to make is that corporations are people as a result of CU and so other human rights apply to them. This is backwards. SCOTUS and lower courts basically established that free speech applies to corporations same as individuals. But it did not establish their personhood. This is exactly equivalent to saying that a corporation has to pay taxes like a person. It does not make it a person.
So what people get wrong is they say “if a corporation is a person then it gets to do X”. Thats incorrect, nobody except talking heads on TV called it a person. Similarly “if a corporation has the right to free speech it has the right to do X” is incorrect. Having one right does not confer all rights. Again think of it as the idea of corporations get to pay taxes. People get to pay taxes. This did not make corporations people and did not confer any other rights onto corporations.
Ugh. CU doesn’t state that corporations are people. They can’t vote or own guns or get married or divorced. They can’t be legal guardians to children or pay income tax. They aren’t entitled to Social Security benefits and their coverage for health insurance may be denied for preexisting conditions. What CU said is that collectively people can use company resources to exercise their right to free speech and established the concept of super PACs.
This isn’t to say that it was the right decision (certainly seems to have done some very bad things). But “corporations are people” is a lay person talking point, not an actual legal doctrine. Therefore you can’t just apply it to other cases because there is nothing to apply.
You are correct that free speech isn’t limited by your citizenship status.
Isn’t Alphabet and other tech companies technically Irish owned? Doesn’t Saudi Arabia own a chunk of Twitter? Seemed like the whole ownership ship justification is a cheap canard.
Sure but even the supreme court disagrees with the supreme court. Treating their rulings as the best or canonical interpretation of a case doesn't make much sense.
It's not like any interpretation is valid but there are plenty of valid ones.
By definition, the Supreme Court's decision _is_ the canonical interpretation. Whether you disagree with the decision has no bearing on the matter.
And of course it makes sense, because the legal system was created by the very laws it upholds. If you think it should be different, then you'll have to convince a lot of people to change a lot of laws and probably parts of the US constitution
This is not affecting US citizens' legal free speech rights. You have the right to say what you want; you don't have the right to say it on a specific platform. You had free speech without TikTok before it existed, and you'll have the same amount of free speech if it does not exist again.
This is exactly the simplistic framing the person you replied to is talking about. So let's take an absurd extreme. The government designates a 1x1 mile "free speech zone" in the middle of Wyoming and says you're not allowed to speak anywhere else. You have the same amount of free speech as you did before, right?
Another extreme, let's say the government declared that you may speak freely but only by filling out a web form routed to Dave. Great guy. I mean they haven't technically taken away your right to speak? And someone will hear what you say.
Both of these would he flagrant violations of 1A as I'm sure you'd agree. But what this means is that implicit to 1A the government has limits on how many places it can deny you speech and limits on how much they can deny you an audience. And you can't hide behind the "well it's just divestiture not a ban" because the courts aren't blind to POSIWID.
So the more nuanced question is does banning TikTok meaningfully affect the ability of Americans to speak. And I think because of how large they are you could answer yes to this question. Americans know exactly what they're signing up for with their TT accounts and want to post there. TikTok but owned by an American would be legal so the platform itself isn't the issue. And saying TT can't operate in the US and actively preventing Americans from accessing it are two very different actions.
>Another extreme, let's say the government declared that you may speak freely but only by filling out a web form routed to Dave. Great guy. I mean they haven't technically taken away your right to speak? And someone will hear what you say.
This argument touches on the more valid defense for TikTok: restricting which people can host speech is a good way to restrict content, by punishing those who tend who host certain kinds of content. Personally, I'm okay with requiring a US company control TikTok in the US for national security reasons, but I would've preferred the law go through strict scrutiny. Laws can restrict what would usually be Constitutionally protected rights as long as they have good reasons and little room for collateral damage. If what Congress has been claiming is true, this law should pass that standard.
Actually, both of those examples might be legal (assuming the form is applying for a permit for some specific event/location). Time, place, and manner restrictions have long been upheld by the courts. What isn’t legal, or at least what requires strict scrutiny, are content restrictions.
Cool, so is all US companies in all other countries around the world then, no protections. All countries in the world, USA just showed it is perfectly fine to steal a foreign companies' asset. Let's do that to all USA companies, Apple, Amazon, Nvidia, Tesla, Boeing, Qualcomn, Intel, all of them. U know how rich you will be if you just got a piece of them? U know you could end homelessness, poverty, balance trade, stabilize your currency, elevate tax revenues, get free education and health care for your citizens, provide great jobs if you just got a piece of USA companies? Now you can! All of them can be Indian, Germany, France, UK, Poland, Brazilian, Mexican, Canadian, Kenyan, Egyptian companies. Everyone gets a piece, everyone gets them equally, everyone will benefit and be happy!
I get this argument, and obviously it's not stopping people from uploading the same content other places. But isn't there (or shouldn't there be) something about not banning what people can consume? Like could the US ban aljazeera? Or banning foreign books?! And still TikTok is different, because it's about the potential for quietly manipulating or curating what is seen, even if that content is produced domestically... And even if people can use other apps, there's still a community and subcultures that are being dismantled.
