40four

Okay, first off, without honestly reading the whole article (I tried but I just don’t have the patience), a quick red flag is the final analysis is only through Fable? Isn’t that inherently going to introduce unwanted bias?

Whatever. Doesn’t really matter much. My next thought is, I get that requiring an SVG is adding an extra layer of complexity as far as the art goes, but why is nobody talking about that the actual art is absolute trash?

I get it. It’s basically a meme at this point and it’s a fun game to play with the models. But my thought is it should be illuminating to anyone who is an artist that LLMs are still a long way off from taking your job :)

You’re really going down a completely different rabbit hole here, that has nothing to do with with the technical content of the post.

I’m not saying you’re right or wrong. I’m saying you aren’t adding anything to the discussion of the technology itself presented in the write up.

I took the bait too and responded to this post, but it’s the third “green” screen name I’ve seen in the last couple days with the same text in the bio:

“When dang and tomhow ban all of us humans, only bots will remain”

So I’m afraid we’re responding to a bot face palm

If that’s what you took away from this, I’m afraid you missed the point. The question isn’t “will anyone actually play this?” No, they won’t.

It’s a demonstration of running web technologies (on the front end, and rust on the back) on a device that was never meant to accept them. And to run something outside of the assumed capabilities of the hardware for that era. I found this very interesting to be honest

I like this distinction “dirt notebook”. As someone who was never naturally good at taking notes, I’ve only later in life made it a consistent habit. That being said I’m realizing all of my notebooks of “dirt” notebooks. There’s no rhyme or reason to it. No consistent format. Just pure stream of consciousness trying to capture whatever was in my head at the moment, or whatever I’m trying to remember from a meeting or call or whatever.

The charts are certainly overwhelming. I had to tap out half way through maybe

I’ve been using Firefox on mobile (and desktop) for years. I still don’t understand all the hate thrown at them. Whatever the downsides/ shortcomings people see, they are irrelevant to me, because I hate Google more & refuse to give them my data. They are good browsers & work just a well as Chrome 99.9% of the time for me

I just mean no matter how hard anyone tries, I don’t see how useful these systems would be in practice. Sure they demonstrated a “working” system. Plenty companies sell products that “work” to one extent or another.

But how useful is it really to get a result of “This is 80% likely chance of being LLM generated”? Or 75%, or 95%? What if the text is a mix of human written text and LLM text? How would you even begin to test that?

I suppose a text that is half human half LLM would theoretically score in the 50% range, but do you see the problem? You can slap a confidence % score on a test run, but interpreting the results leads to a whole other can of worms.

Point is there are so many variables, and it’s not clear that the result from any of the systems is even valid or applicable to help you make a decision in a real life situation.

I could be wrong, but I just don’t see how trying to “detect” LLM generated texts is ever going to work. The only thing that makes any sense if you truly want to have confidence a human wrote it is some type of “proof of work“ system. I think there’s a lot of interesting ways to approach the proof of work problem with different pros and cons, but that is where our energy should be focused if we seriously want to solve this problem.

That actually makes a lot of sense, and it helps me wrap my head around why the technique exists at all. As someone who didn’t get into web programming until 2015 or so, I didn’t quite understand at first the usefulness of this. But for sites like this built in the 90’s it was a totally different world

If I hear somebody say “deterministic“ one more time I might lose it lol :)

Okay I guess? So then you politely answer to them the “app” is their web browser of choice! Problem solved

I actually kind of liked the timer. It gave me a sense of urgency and exhilaration! Word games have never been my favorite, but this has a different feel. But I can see giving the option to turn it off, and fine tuning the timed mode. Seems like some real potential here

Yeah, I was lucky. My Samsung device is officially supported. As you said, since no one else has had the interest to do it for your Nokia, it’s basically up to you to make it work, since they rely on volunteers.

I wouldn’t begin to know how to take the base ROM and adapt it to a specific device, but I’d imagine any of the frontier LLMs could probably make quick work of it. It might be worth taking a weekend, and spending some tokens on coaching a model through the task?

I don’t disagree with what you said, but that same logic applies to every frontier model, or even lower tier models. You are subject to their creators bias whether you like it or not.

So what you are really saying is that you don’t accept SpaceXAI’s bias, and you’ll plant your flag elsewhere. It’s not that the other camps don’t have their own bias.

I don’t mean to be rude, but did you with a straight face say the Chinese models “aren’t obviously active about their politics”?

Ask one of those models a few critical questions about the CCP and Chinese history and see what kind of results you get :)

It’s too close to China & Russia, whereas Venezuela had nobody in the vicinity that could help respond.

Recently installed LineageOS on an old Samsung tablet I had laying around. The thing was barely usable with whatever the highest Android version was it was comparable with. Now it works wonderfully! Really happy with it. Privacy features aside, Lineage can do wonders to revive old hardware

I don’t know that there’s anything wrong with the way the rule is written, the issue here is the way VAR was applied. VAR is only supposed to intervene when there is a “Clear and obvious error.”

Balugun’s play certainly “endangered the safety of an opponent”, as the red card rule reads. Intent doesn’t matter as far as the rule goes. But the call on the field will always be subject to the referee’s judgement on the field. They are weighing a variety of factors, and intent plays into that judgment I think.

Bologun’s challenge was certainly red card “worthy”, but I think most people agree that the initial yellow card was the right call, especially since it wasn’t intentional. The ref saw it full speed, made his judgement, and that should have been the end of it.

VAR likely overstepped their mandate here asking for the replay review. I don’t think that was a “Clear and obvious error.” so they influenced the ref to upgrade to red. It’s especially upsetting when there are many other glaring examples of yellow cards in the same tournament that they did not send to review.

Interesting, thanks. I admittedly spent zero time looking into it :)

I’m surprised open source contributions count for so much. first I thought was “is that something people actually list in as resume?”. But it looks like it pulls your GitHub account and appends that information.

That kind of unfortunate for anyone who doesn’t use GitHub

You’re making a good point. I don’t disagree with what you’re saying. But I think my point got lost.

I don’t agree with “Software development is where money is made for these labs”. Coders will inevitably eat up the most tokens & buy the bigger $200 subscriptions because we want to keep working.

But us coders are still the small minority of users. They aren’t counting on us to get to trillion dollar evaluations.

They are counting on the regular folks to buy the $20/ month subscription. It’s really easy to run out your free tier usage these days, asking questions that have nothing to do with coding.

So my point is what does that output look like for someone asking a question about politics or world news?

“I'm a little confused, is this an ATS system that anyone actually uses?”

You read my mind. If the answer is “no”, then we can ignore this.

It’s hard to argue against the open weight models if your only concern is coding. Which, for many of us hackers here in this forum, it is.

But I would like to point out that the overwhelming majority of people using LLMs aren’t programmers, don’t care about coding, and couldn’t even be bothered to “vibe code”.

So we should consider the bias of the output of these open weight models, and what that looks like, outside of the context of writing code.

Yeah exactly. Drones have changed everything. What might have used to be routine is now no longer the case. Iran just has to project power in the media, and shoot some drones out at some boats every now and then and they effectively “control“ the straight. The shipping insurance markets skyrocketed as a result, so ships literally couldn’t afford to take the risk and stayed put.

Yeah that’s a good point. I don’t disagree with that. I just think the US administration’s strategy is something totally different than anyone is talking about.

I don’t think it’s a coincidence US aggression towards Venezuela, Cuba & Iran are all happening at the same time. These things are all connected and nobody’s really talking about it.

I don’t think regime change was the strategy, I think they were happy with just a “reset” of the top leaders, same as in Venezuela.

If they did get full regime change it would have just been a side effect. They were hoping the Iranians would seize the opportunity and rise up, but that didn’t happen. And I don’t blame them, that’s a big ask when they are getting gunned down in the streets or executed for dissent on a daily basis.

They were probably also riding off the “high” of how shockingly easy Venezuela went. But Iran is much more complex obviously

Controlling the straight was never one of their objectives. Decapitating the regime and degrading their military capabilities was the primary objective.

As others have said, the US can “reopen” the straight at any time they want. It’s not an issue of capabilities. But it’s very resource intensive and very expensive.

The logistics of escorting ships in and out of the straight isn’t trivial. I forget the name of the operation, but they did implement it for a few days before shutting it down. Politically, I imagine it’s pretty hard to justify the cost /benefit

On the Iranian side it takes a very small amount of resources and logistics. All they have to do is project power, whether they have it or not, and the shipping & insurance industries have to respect it.

Drones are really cheap, and that’s about all it takes for Iran to leverage their influence over the straight. Which is kind of crazy when you think about it. But it’s about the only bargaining chip they have left and they aren’t going let go of it easily.

I’ve been experimenting with a pet project that aims to solve this problem. “AI detectors” are certainly unreliable, I’m not sure they’ll ever get to a state where you can trust them.

I think concepts like this are the only reliable way to prove something was written by a human. A full replay like this is one way to do it. I think there are some other feasible ways to achieve this, maybe in combination with a full “replay”, but some sort of “proof of work” is the way to go I believe. As LLMs become more ubiquitous, I imagine products that solve the problem can be a real business opportunity.

Hey I hear you, I’m not trying to make this a political argument of who’s dropping bombs on who, or the American government is better than or worse than the Chinese. But what I said is a matter of fact.

We can debate the semantics of whether “created by” or “subject to” means the same thing in regards to the Chinese government, but that is neither here nor there.

I’m happy to take your wording that they are obviously “subject to” the Chinese government. That logically means they are subject to carrying out the CCP’s long term strategy. And as you said “whose whims may change at any given moment”.

That directly relates to the OP’s fears, that these models could be taken away at any given moment. “The spigot can be turned off at any time” as they put it.

Or another possibility is they will never turn the spigot off, but they will engineer it in a way to best achieve their goals. My bet is that’s the more likely outcome.

I simply disagree with the OP’s description of the problem as “open weights models are the result of philanthropy by some private org”, I think the problem is much more complicated than that

I don’t think anyone seriously believes any of the Chinese models are ever going to “overtake” the American frontier models. I doubt that that’s even their goal.

But if they can stay on pace, within say 6 to 12 months of the bleeding edge of the American frontier models, that’s a huge problem.

If they can just piggyback on the Herculean efforts of Anthropic, OpenAI, Google etc., accept a little bit of lag, and save billions of dollars? Why wouldn’t they?

And for the end user, why would they pay a premium subscription price for something they can just wait six months for and run on their own hardware at home? In my opinion, this is the cat and mouse game that’s being played right now. And I suspect it’s intentional on the side of the open weight models. I would bet they are playing a war of attrition

We should address the elephant in the room. The problem with the future of open weight models is not they are created as a result of philanthropy by some private org. All of the top contenders are created by the Chinese government.

I don’t think we should describe these companies as simply releasing these highly capable open weight models out of the goodness of their hearts