Ohad Asor

@Lee715495232368 @newstart_2024 1000-step solution is nothing, he has no idea what he's talking about. Yes LLM might be able to make 1000 steps but that's not nearly enough. Logical solvers can already do millions of steps *per second*.

@newstart_2024 @ohadasor Thoughts ?

Former Google CEO Eric Schmidt drops a chilling warning on AI's future "Within 5 years, AI could handle infinite context, chain-of-thought reasoning for 1000-step solutions, and millions of agents working together. Eventually, they'll develop their own language... and we won't understand what they're doing." His final words: "Pull the plug." This is the man who ran Google talking about the singularity. 2:59 clip inside—must-watch.

@journalistica_ Sure, I spoke about it in length here youtube.com/watch?v=JVLpxm…

@ohadasor You don't think that even with their inherent approximate nature (never being computationally correct by design), LLMs can still contribute to the progression of many important fields (even if this is just the middle point before integrating Logical systems)?

This tweet is +1y old, but I've said it for much longer. ChatGPT came years ago, and "AI" is still more or less the same. Against all "expert" predictions, it didn't cure cancer, it didn't colonize Mars, it didn't discover new physics, it didn't really improve much. Because I was right all along: machine learning is around its peak, and it is a huge bubble. Nowadays it's obvious, but back then I was a single voice against the whole world. Not only that machine learning is fundamentally incapable of logical reasoning, but the architecture of those giant data centers, is also incapable of logical reasoning. And I keep telling you: the real deal is Logical AI, and we are the leaders of this segment. Machine learning is only for translation.

@Krawlarr We did launch a version of the logical AI engine. And testnet is around the corner. This post is about AI. It is clear that you are a troll paid by the scared "competition".

@ohadasor Yes, but you've not launched yet. Not even testnet. How do we know if Taunet works?

The path to AGI is in new methods regarding Atomless Boolean Algebras. And those are all patented. So you are stuck with me, sorry.

Really really funny. Not even close. Maybe his analogy holds for extremely simple logical systems, far from enough for software specification. He thinks he's the first to find the connection between Datalog (polynomial time) and matrix multiplication (while seemingly he missed expressing recursion using matrix inverse). Can kill his argument using undergrad complexity theory!

@Freshbt2 Atomless Boolean algebras are much beyond just 0s & 1s. For another thing, Tarski's result of course holds, but we were able to achieve the goal by relaxing his assumptions and weakening his requirements, in a way that is good enough for practical applications.

@ohadasor So the Tarski's Undefinability of Truth has been solved, and no one bats an eye. (Describes 0s & 1s with 0s & 1s) Hats off 🎩

@arnau00296942 It can prove anything as long as you express it in the Tau language. This is already a very big limitation, there are only certain things that the language supports. For example, it cannot express problems that require above double exponential space to solve.

@ohadasor The little bit I understand about Tau Language I find fascinating. I don't understand it's limitations though, can it for instance prove theorems from specifications? What can't it do provided you can specify it (that a human can)?

Asked ChatGPT "why can't machine learning perform logical reasoning?" and got a good answer! Too long to paste here, so try it yourself

Machine learning models / LLMs excel at patterns but will never offer logical correctness for non-trivial/complex problems. I'm excited about formal software synthesis from logical requirements, where correctness is guaranteed by construction rather than hoped for.

Watch my interview with @ohadasor from @Tau_Net here on MLST - youtube.com/watch?v=JVLpxm…

Imagine software that adapts to you—individualized to your needs. Ohad Asor and the Tau Team are building the first and ONLY blockchain that its users fully control: -- $AGRS

Light years away, not even close and doesn't even say anything. LLM can't even start helping to code the Tau language. x.com/i/grok/share/t…

@P33RL3SS @f4talStrategies this conversation was a nice demonstration of the original post

@f4talStrategies @ohadasor When you need certainty, when you want total accuracy, reliability, you benefit from using a machine. Whether it be a calculator, or a computer. The problem with LLM outputs is the lack of accuracy in the results.

Quantum mechanics works very well (except when it doesn't), but whenever a quantum theory is illogical, they say "find a better logic". LLM works very well (except when it doesn't), but whenever it is illogical, they say "what is logic? there isn't such a thing". Coincidence? No. Human nature, yes. Same human nature that made humanity susceptible to religion.

@xmr_chan Morality is subjective, yes, but this doesn't make it any less important.

@xmr_chan The alternative is morality which will indeed prevent us form killing.

@Rimka1000 @TauLogicAI We have concrete release, I don't know why you say that

@ohadasor @TauLogicAI ok but promises only bind those who make them...people have been waiting for something concrete for years and so far nothing…be concrete please

Show me one blockchain project which is not just more of the same thing. Show me one AI project which is not just more of the same thing. I'll show you one that ticks both boxes: @TauLogicAI $AGRS. Read my two recent articles here to see why it's the case

Tried (link below), at the end it admits that no language can do it, but adds "practical implementations can indeed manage safety or command filtering based on pre-defined criteria" which is just a joke: pre-defined criteria? For one thing, you want to give criteria to the system on-the-fly, you don't want only a fixed small set of predefined yes/no or numeric parameters that are supposed to encapsulate the concept of safety (as one example). For another thing, that'll require that the command explicitly touches those parameters, so the safety check may still be bypassed by modifying the parameters implicitly. It simply ignored the key word here which is "contradicts" and wants you to tell the computer yourself (by explicitly pointing to parameters) where the contradiction is. x.com/i/grok/share/Y…

@ohadasor Hi, not really a programmer so I asked a q here, what’s your response? x.com/i/grok/share/m…