Ero

Depends on the strain, but it generally makes everything 'softer'. Slows you down. Music sounds so good when high.

Hah, no bias here. Hornets are necessary for the ecological balance of insects. Got a tattoo of one. Might get a centipede soon.

Just a recent example - when the Oracle deal for the Memphis Datacenter broke down, he instead decided to build it himself. In 19 days, he now has the most powerful computing cluster in the world, sitting at around 2.5 exaflops.

Hah, I can. I am sure you are responsible, I am not saying otherwise, but I am just giving a little pushback so you don't fall prey to the current FOMO which will undeniably wipe out a lot of wealth/investments if you are not careful. Good intuition. A technological innovation is normally an agglomeration of separate S-shaped curves that build on top each other until all the vertical and horizontal growth is exhausted. AI is supposedly different due to the fundamentally different end-case scenario, which is the so-called 'superintelligence explosion'. We are undeniably headed there, however, imo it will take a lot longer than Musk or Leopold are currently predicting. Absolutely. AI has a lot of room to grow, but for that to happen, LLMs and Transformers need to stop sucking all the air (talent and investments) out of the room, which is why I think a bursting of the bubble will be useful in the long run.

The IMO result is extremely impressive. I am waiting for the paper to dig a little deeper into their exact method, outside of what we already know about the Lean-based RL. This is an interesting inquiry. Having worked with Lean in academic setting, my general impression is similar to that of the following paper about the proof as a social process. There is a disconnect between the human-centric proofs we write and the computationally verified proofs by Lean. The transfer in both directions is non-trivial, and I wonder if at a certain scale it would be possible. (I can see for example from the AlphaProof IMO files that the annotations are human-made) Most of the times the approach/ framework used is far more important than the proof itself, as was the case for example with Grothendieck's etale cohomology, which had an impact far wider than just its use for the Weil conjectures. I am not yet certain myself how much of that can be supplemented by formal verification tools. Lean itself is also very definition-dependent and sometimes it is precisely the reworking of the definitions that is needed, instead a procedural generation based on it as an axiom.

Fif is on a different level. His trolling also deserves an award lol.

Thank you! One thing that helped me is that I hid my debit/credit cards from myself (I don't use cash). The smoke shops didn't accept iPhone tap to pay, so it created enough distance/resistance to be able to realize and stop myself.

My personal 'cocktail' was Caffeine, Alpha-GPC, Omega-3, Lion's Mane and Gingko Biloba. It definitely helped with the consistent 10-12h/day of math lol.

That's such a funny thing to say. I drink like half a liter the moment I wake up. Never thought I was addicted hah Black coffee smells and tastes so good.

Do you plan on opening straight from the start with 'You are God' or are you planning to warm him up first lol

Predicting outputs from inputs is what all ML models do (that's why they are called models). They need this much compute, because they are 'searching' a parametric space of trillion parameters - it comes from the sheer size of the models. You are intuition is on point IMO. Similar to Gary Marcus, and more recently Yann LeCunn, I am a proponent of the idea of 'embedding' the representations of human thinking/ knowledge to achieve immense efficiency speed-ups. I am quoting the Lie-Poisson paper in my previous post for this: 'The advantage of PINNs is their computational efficiency: speedups more than 100,00x for evaluations of solutions of complex systems like weather have been reported' - referencing Accurate medium- range global weather forecasting with 3d neural networks and Fourcastnet: A global data-driven high-resolution weather model using adaptive Fourier neural operators

How could I? I am writing a thesis, referencing many of those papers. Nowhere did I say it is the end of the process. GPUs will always help, but what I was pointing at is that alone won't be enough. There are physics- and biology-related systems that are provably unsolvable by Transformers (part of what I am working for my thesis). That is what I meant by that not being sufficient, no matter how much compute you throw at it. Again, you seem to take for granted 'new paradigms' coming along. As someone who works on the categorical/reasoning side of AI, I can tell you that their construction would be painstakingly slow, utilizing all the domain-specific knowledge we have. An example are Lie-Poisson Networks, part of a broader class of PINNs. Embodied AI would definitely be a revolutionary technology, but you won't achieve it without having first some breakthroughs in the PINNs and Reasoning paradigms. I agree with the AI ecosystem vision, moreso than singular foundational models. Local domain-optimized models would be far more effective for the wide range of scenarios that would cause the disruption you are foreseeing. The thing is, when you say exponential, then you assume a multiplicative improvement year-by-year, similar to the AGI by 2027 wacko, whereas I am predicting more a cycle-like trend with general improvements with two important notes: first, the current peak is not the final, and second, the 'intelligence explosion'/ 'run off' scenario is almost impossible, considering the underlying energy constraints - with the current Energy production, there is just not enough to handle such a scenario (Wells Fargo predicts a 550% surge by 2026 to 52 TWh). We gotta scale up 5-10x Nuclear and work towards Fusion to have even a remote chance of satiating the upcoming energy demand.

Absolutely. I agree with your frame about evading politics, I don't see how that would be productive.

Not within the current paradigm, i.e Transformer/LLM . Mira Murati, the CTO of OpenAI has already stated they don't have anything in production that is much better. Check the MMLU benchmark plateau in the second image - you can see only marginal improvements. LLMs are already at the their maximum, since they have exhausted pretty much all data, synthetic data doesn't work as yesterday's article from nature shows. What you are referring to is ASICs, or dedicated chips, such as those my classmates built for etched.ai. The problem with them is that they work only for inference and not for training. You still need GPUs for that and even with the latest H100, LLama 3 or GPT-4 is estimated to have exceeded $1B, and that is only training. So no, dedicated silicon will not solve this problem. I agree with this, but you seem to take for granted the invention of 'new more effective techniques'. The last innovation was the transformer and that was 7 years ago. We haven't had any vertical/ architectural jumps since then, it has only been scaling. And even that doesn't prevent representational collapse or incorrect reasoning.

Hows your reading comprehension? Let me do a breakdown of what you said, vs. what I said: consciousness will be heavier, anchored, simpler vs. fluid, I have had vivid dreams and spiritual experiences they will be overtly much more interested in playing and sleeping vs. mnot interested in playing or sleeping, if... interested in 'abstract stuff' will be very pragmatic and without qualms, openly selfish and in search of a "simple" paradigm vs. I study Pure Mathematics, which is as abstract as it gets (which is also as far from pragmatic as a subject can be), have a disdain for reductive and 'simple' paradigms If you really go against his interests, he can on the other hand go much further from time to time in physical violence. vs. have NEVER resorted to physical violence OUTSIDE when it has been absolutely necessary for my own and my friends' safety.

Regretfully, I this is how it had to be for me. I was smoking/ taking edibles almost daily to the point that it started directly interfering with my responsibilities for uni, so I had to a cold turkey stop. I haven't smoked in like 7 months.

@Schizophonia I have very very high levels of Testosterone naturally and almost entirely disagree with your following description: My consciousness is very fluid, I have had vivid dreams and spiritual experiences since very young. I need 3 sec to begin experiencing the 'Dancing Buddha'. I am not interested in playing or sleeping, I study Pure Mathematics, which is as abstract as it gets, have a disdain for reductive and 'simple paradigms' and have NEVER resorted to physical violence outside of being jumped or surrounded on the streets. Be careful with being this reductive or simplistic in your assumptions.

@cjoseph90 Have you been experiencing any long-term side effects? I am asking, because I was really big into tea, but when I started doing a lot of cognitive work for uni, the caffeine was just not enough. I started drinking like 3-4 cups a day for the last year or two and haven't had any problems with sleep or heart rate. I am physically active, so that might be helping, but I wonder if I should dial it back.

@RedLine That is very interesting. Mushrooms and LSD are very different for me, Haven't tried DMT yet, trying to get my hands on it. From my experience Mushrooms are very magical, almost archetypal in their imagery and messages. I have had hallucinations of serpents, pyramids, my ancestors, whilst simultaneously the experience being very blissful. My first mushroom trip was 2g, which I barely remember, same as the 5 g I did a year later. All my other mushroom trips were in the 1.5-3 g range with extremely positive impact, meta-awareness, ego death and the sort. I They feel good only when I am in nature. LSD I have probably around 20 trips, the lower doses of which are 150-200 ug and the higher doses being 250-350 ug. The lower doses are perfectly manageable, with great insights into cognition, science and technology. LSD is very cerebral and sharp for me. The sharpness gets stronger towards the higher doses. For some reason my last 4-5 trips on LSD were all chaos-like. For the first of them I made the mistake of having watched Horror (Evil Dead Rise) and researching the sigil Solomon mythos, which backfired. The subsequent ones were very much geared towards the realizations of the sheer Chaos in Consciousness. The analogy I have is that most of the nature documentaries show only the butterflies, colorful birds and puffy animals, but don't show the parasites, hornets and all the murder and death that is part of the cycle of life, the latter of which was essentially where my trips were forcing me to look at. Wouldn't call any of the latter trips 'bad', but they were definitely humbling. So much so that I haven't tripped on LSD in almost a year and I am lowkey anxious to take the one tab I got rn.

Elon has definitely done some questionable 'marketing'. His famous predictions that never happen are part of it. But if you look in general at Silicon Valley, it seems to breed that type of over-zealous vision crafting. Sam Altman, who in my opinion is straight-up a sociopath, is currently doing some of the most bonkers predictions, all to rile up investors. Amongst their ilk, Musk is probably the one who deserves the most credit, but that still comes with a disclaimer about his political opinions and employer style. I have a friend who worked at SpaceX and he told me that with Musk at the helm, you can kiss goodbye your personal life.

I read on the author's wiki, that his original work 'All Yesterdays' was based on an alternative evolutionary history of cryptds. There is this awesome series on YouTube, covering the latter. Could be an interesting watch:

I loved this video when it came out. I have a sweet spot for this type of sci-fi/ cosmic horror. I think the author is fundraising to complete the translation (link)

This is just False. However, it is true that we are heading into an AI winter, since the current scaling paradigm has already reached the two fundamental bottlenecks of data and compute/energy. I think it is necessary to sober up. I am foreseeing a year or two until the markets correct for all the burned money that hasn't seen returns (check out Sequoia's Article on AI's 600B Questions) We need fundamentally different approaches to architectures, employing all the scientific knowledge we have through a functorial bridge that is yet to be built (Paper by Google Deep Mind and others suggesting precisely that)

Looking forward to that one.

@Phil King Ah, I forgot about that one, thanks. Imma rewatch it. I have had far worse, trust me.