How about if you run this loop (one year from now) on this kind of hardware but with something like Claude/Kimi K2. How about that? Because that's where it'll go.
Why so negative lol. The speed and very reduced power use of this thing are nothing to be sneezed at. I mean, hardware accelerated LLMs are a huge step forward. But yeah, this is a proof of concept, basically. I wouldn't be surprised if the size factor and the power use go down even more, and that we'll start seeing stuff like this in all kinds of hardware. It's an enabler.
You don't know. You just have marketing materials, not independent analysis. Maybe it actually takes 2 years to design and manufacture the hardware, so anything that comes out will be badly out of date. Wouldn't be the first time someone lied. A good demo backed by millions of dollars should not allow such doubts.
Did you not see the chatbot they posted online (https://chatjimmy.ai/)? That thing is near instantaneous, it's all the proof you need that this is real.
And if the hardware is real and functional, as you can independently verify by chatting with that thing, how much more effort would it be to etch more recent models?
The real question is of course: what about LARGER models? I'm assuming you can apply some of the existing LLM inference parallelization techniques and split the workload over multiple cards. Some of the 32B models are plenty powerful.
What the US media (and Elon Musk) call EU censorship is actually a request to follow EU rules if they want to operate in the EU market. What, exactly, is controversial about that?
Yes? It’s not exactly a surprise that you are expected to follow the laws where you do business. This doesn’t mean those laws are inherently good or bad, that’s a judgement which requires analysis to make and businesses quite reasonably might choose not to stay in a market based on that decision as Google did with China.
It's not that controversial, every single country has limits on speech, including the US. So European countries control a little bit more than the US, largely when it comes to racial abuse and other hate speech. So? The American model when combined with social media and the internet appears to have disastrous outcomes, judging by who has been elected there. It clearly worked in the past, but not any more.
Americans supposedly being outraged at other free, democratic countries (often in reality both more free and democratic than the US) having different laws regarding speech is really just a smoke screen for what they really want: for their social media companies and billionaires to completely control our media, so that we end up just as fucked up and insane as they are. In the end if we allow Americans to poison our countries, we will lose our freedoms and democracies. Why would we allow that? What do you expect?
P.S. it's cringe to cry about lack of free speech in Europe as if we've changed. We never, ever had 100% free speech in Europe. Stop trying to hark back to some free speech utopia that literally never existed. This is the continent that up until 110 years ago was overwhelmingly ruled by kings and queens and indeed we are in many ways far more conservative than you are. Get over it and stop trying to turn us into you.
I would just like the early American project of liberal democracy and Constitutional rights to outlive American capitalism and American militarism, even if it means it survives it in some other country. Because it's looking pretty bleak over here.
We ought to avoid repeating your mistakes, no? Maybe unlimited campaign donations and so on, all this wonderful "American free speech (money = speech)" is a fundamentally bad idea. Worked exceedingly well for ~225 years, now it has lead to the implosion of the empire by electing a sociopathic retard to the presidency. Yes to free speech, no to whatever fucked up shit the US, its billionaire "libertarians" and Christian nationalists are pushing for us to adopt here in Europe.
If the likes of JD Vance are pushing for us to adopt his idea of free speech, you can be sure it's a bad idea.
The American political system didn't implode until its system of capitalism had the conditions necessary to escape its popular control. That wasn't necessarily an eventuality. We had a semi-functional campaign finance system in living memory.
Without the protections the Americans tried to shove into the First Amendment (which did not include anything about corporations at the time, as they did not exist) being enshrined into law, I worry that your issues with capital-government overreach will arise even faster than ours.
I don’t disagree with you but I disagree on a point of history.
> Without the protections the Americans tried to shove into the First Amendment (which did not include anything about corporations at the time, as they did not exist) being enshrined into law
If I recall correctly, Britain had joint stock companies from the 1600s, and Adam Smith and all that. They also even before this had “trusts” and “trusts which own trusts” which had certain rights, and the court of chancery had established precedent around these.
The French also had a massive state stock company in this time, and it became a massive bubble which imploded in XXXX. This attracted a lot of attention and commentary and it’s impossible that the American Founders were ignorant.
The Brit’s never had a freedom of speech, but in English common law, companies had property rights, standing to sue, and so on. Most activities a business person could take, they could take on behalf of their company instead.
So in the American context, it seems that the founders were likely aware of corporations. Why they didn’t put explicit limits in the first amendment, who knows. Maybe it just didn’t seem important at the time.
It came very close in the 1930s, it is arguable that the New Deal headed off revolution
The USA should have been considered a pariah state since the Gulf of Tonkin incident in 1964, now it is rapidly becoming one
The USAnian system has been a corrupt oligarchy with only trappings of democracy since it's inception. Those "trappings" run deep, but are not allowed to unseat the true source of power: money
The Founding Fathers [sic] gave that a lot of thought and worked very hard to make it that way from the very beginning
Ask Alex Jones about his free speech on Sandy Hook to understand how bad (EU) censorship really is!
Jokes aside. Restriction of freedoms, including speech, is not bad by definition, it's the scale and intention behind it that matters but this aspect is always missing, kind of censored, in public debate. You may downvote me now :-)
Edit: In the same sense, Alon does not cry about specific and obviously unjustified cases of EU censorship on X.
Of course not. It's only censorship if the rules are censoring rules. Just because a billionaire right wing extremist cries "cEnSoRsHiP" everytime people who criticise him aren't imprisoned doesn't mean it is.
The reason is that the post you link to is overly simplistic. The only reason why Simon's experiment works is because there is a pre-existing language agnostic testing framework of 9000 tests that the agent can hold itself accountable to. Additionally, there is a pre-existing API design that it can reuse/reappropriate.
These two preconditions don't generally apply to software projects. Most of the time there are vague, underspecified, frequently changing requirements, no test suite, and no API design.
If all projects came with 9000 pre-existing tests and fleshed-out API, then sure, the article you linked to could be correct. But that's not really the case.
If you start with some working software, you could make an LLM generate a lot of tests for the existing functionality and ensure they pass against the existing software and have excellent test coverage. Generating tests and specifications from existing software is relatively easy. It's very tedious to do manually but LLMs excel at that type of job.
Once you have that, you port over the tests to a new language and generate an implementation that passes all those tests. You might want to do some reviews of the tests but it's a good approach. It will likely result in bug for bug compatible software.
Where it gets interesting is figuring out what to do with all the bugs you might find along the way.
> pre-existing language agnostic testing framework of 9000 tests
if there exists a language specific test harness, you can ask the LLMs to port it before porting the project itself.
if it doesn't, you can ask the LLM to build one first, for the original project, according to specs.
if there are no specs, you can ask the LLM to write the specs according to the available docs.
if there are no docs, you can ask the LLM to write them.
if all the above sounds ridiculous, I agree. it's also effective - go try it.
(if there is no source, you can attempt to decompile the binaries. this is hard, but LLMs can use ghidra, too. this is probably unreasonable and ineffective today, though.)
reply