Alejandro Marin

@DrJimFan Your post make me think of Sora as a verisimilitude engine

If you think OpenAI Sora is a creative toy like DALLE, ... think again. Sora is a data-driven physics engine. It is a simulation of many worlds, real or fantastical. The simulator learns intricate rendering, "intuitive" physics, long-horizon reasoning, and semantic grounding, all by some denoising and gradient maths. I won't be surprised if Sora is trained on lots of synthetic data using Unreal Engine 5. It has to be! Let's breakdown the following video. Prompt: "Photorealistic closeup video of two pirate ships battling each other as they sail inside a cup of coffee." - The simulator instantiates two exquisite 3D assets: pirate ships with different decorations. Sora has to solve text-to-3D implicitly in its latent space. - The 3D objects are consistently animated as they sail and avoid each other's paths. - Fluid dynamics of the coffee, even the foams that form around the ships. Fluid simulation is an entire sub-field of computer graphics, which traditionally requires very complex algorithms and equations. - Photorealism, almost like rendering with raytracing. - The simulator takes into account the small size of the cup compared to oceans, and applies tilt-shift photography to give a "minuscule" vibe. - The semantics of the scene does not exist in the real world, but the engine still implements the correct physical rules that we expect. Next up: add more modalities and conditioning, then we have a full data-driven UE that will replace all the hand-engineered graphics pipelines. openai.com/sora

@Jayde8700 In the same situation, I’d want the same thing. Seems totally logical to me. #DadsAlsoCare

Does this not strike you as odd? “Canadian Prime Minister Justin Trudeau's children will continue to live with him following the split from his wife Sophie, who has already moved out, the pair announced today. Sophie Gregoire Trudeau has already moved out of the family home to another residence in Ottawa.”

@Plinz Cybernetics, in multiple dimensions: social self-organization, tech as an extensor and regulator, cybernetics as a discipline of governance

I don’t think that a gen pop social media platform should rely on enforcing good behavior by technical means, for the same reason that societies cannot replace governance and policing with robots. Instead, it should support social self organization of users. Tpot is an example.

@goth600 Good old-fashioned cybernetics, at the end of the day

It’s impossible to psy-op me. I’ve been exposed to too much information as a child. I’ve studied too much. I’ve worked too hard. My perception rolls are off the charts. The longer you talk to me, the more you expose yourself. The more energy you spend trying to convince me, the less energy I use believing you. My theory of mind uses hyperdimensional pattern matching to see what you’re hiding and what you’re coping for. Your slightest change in vernacular triggers my mkultra cogsec firing pins. It’s impossible to psy-op me.

@Plinz Knowledge is a market, unfortunately. It’s the same pattern I’ve seen in other domains. And, to this matter, poor cybernetics has been constantly raped by different authors exploiting its basic model through different packages

I feel cheated when an author puts together all the basic ideas of the standard cybernetics curriculum, pretends that they are somehow new or newly assembled, and then claims that this would constitute a model of consciousness; it's more blatant than Friston himself

@Plinz For me, Josha Bach is in the list 🙂 Principles of Synthetic Intelligence made a difference in my perception of AI and AGI

In your view, who are the fifty de facto most influential/inspiring thinkers (historically and current, including adversaries) in the public AGI space? I start off with a few suggestions, please tell me who you'd add in the responses. Aristotle, Ashby, Asimov, Babbage, Bellman, Bengio, Brin, Brooks, Boole, Chomsky, Dreyfus, Egan, Feigenbaum, Frege, Friston, Gates, Gebru, Goertzel, Hassabis, Hebb, Hinton, Hofstadter, Karpathy, Kay, Knuth, Kurzweil, LaMettrie, Leibniz, LeCun, Lem, Lighthill, Marr, McCarthy, McClelland, McCullough, Minsky, Moore, Moravec, Musk, Newell, Ng, Norvig, Page, Prigogyne, Rosenblatt, Schmidhuber, Searle, Sejnowski, Shannon, Simon, Smolin, Solomonoff, Sutton, Sutskever, Tarski, Turing, Vaswani, Vinge, von Neumann, Weizenbaum, Wiener, Wittgenstein, Wolfram, Yudkowky, Zuse

@ataiiam @leecronin The statement “thinking is movement through conceptual space” challenge the definition of both, reasoning and imagination (IMHO)

AI helps us decompose 'intelligence' into distinct capabilities. Turns out _most_ of intelligence is reasoning, not imagination. Seems within grasp of current paradigm. Human-level "creative synthesis" also seems within grasp BUT novel thought does seem out of reach. for now..

If we ever get AGI it will bear no resemblance to current AI tech. Current AI tech is a away of turning data into effective algorithms. AGI will need to go way beyond that and build explanations that are formed in an ‘imagination’.

@AndrewYNg I believe that before entering such a complex conversation we should do some information architecture in the conversation itself. This is a pattern we’ve already learned from decades of design work (BTW, I’m hinting at the fact that you are ultimately proposing a design convo).

I'd like to have a real conversation about whether AI is a risk for human extinction. Honestly, I don't get how AI poses this risk. What are your thoughts? And, who do you think has a thoughtful perspective on how AI poses this risk that I should talk to?

@SignalsNSorcery 😂

Has someone checked in on Zuckerberg after the Apple announcement?

@__drewface Attention is the reading head point at something

what the hell is attention. explain attention to me right now. what is attention.

When perspective changes shapes, understanding can change with it.

@AGI_HeavenHell @ylecun Indeed. And when we speak about exponential growth we fall into the trap of seeing it as a curve (a line), when we should find ways to visualize it in a non-linear way, like dots saturating in a region, multiple lines playing together, or some representation that isn’t linear

Wondering how "runaway AI" can run away if it's been "running into walls" for several years now🤔

@elonmusk Same reason dislexia exists? 🤷🏻

If AGI is almost here, why doesn’t autocorrekt work!?

@_florianmai @Artoftheproblem @pmddomingos @sama Statistics are useful because the predicted pattern either matches reality or is convincing. In other words, statistics doesn’t need to be true in order to be useful.

@Artoftheproblem @pmddomingos @sama Then every statistical predictor always hallucinates. So it's surprising that statistics is useful at all?

common pattern: ChatGPT users demonstrate that they understand the limitations and weaknesses of the product *extremely well*; experts in the field continue to worry it’s just too complicated and the users won’t be able to understand

@_florianmai @Artoftheproblem @pmddomingos @sama Hasn’t Gestalt stated that (with different words) decades ago? 1) Take given patterns 2) Complete as you can

@sama That’s why you usually exclude experts from your user research (unless you are designing for them)

@FelixDelong Makes sense! (IMHO). If nature bets on recursiveness to scale (fractals) then the resulting design system should be high on homoiconicity (the data resembles the code which resembles the expression of the code)

So let me get it straight - DNA is like binary, except with 4 values and proteins are like assembler, except with 20 values. mRNA is a compiler between the two. Not only DNA has a language, but it is also actually language embodied. And a programming language at that 🤔

@leecronin I haven’t read your paper but as a designer I intuir Assembly Theory makes tons of sense if we observe life as a huge design system at play

Assembly theory is going to change things. Since Darwin, we have tried to understand how the countless forms of life can evolve in a universe with fixed laws. The laws of physics are connected to life, evolution, and human culture, but they can't predict their emergence. Assembly theory doesn't change the laws of physics, but it changes how we think about selection. In AT, objects aren't seen as single points but as a collection of possible formation histories. AT demonstrates that objects can show evidence of selection, whether part of an individual or not. This requires the new concept of Assembly, which measures the amount of causation needed to create a group of objects. AT reimagines matter in terms of assembly spaces, offering a way to unite physics and biology in a new field that starts playing a significant role at the chemical scale, where history and selection become important factors.

@designaid_web @zendadddy @oatly In this case, eyeballs on your balls

@zendadddy @oatly I think is brilliant. Marketing is how to get eyeballs on your stuff.

is the marketing team at @oatly okay? fckoatly.com

@SignalsNSorcery I think those who might listen to synthetic music will probably be the same listening to Mood playlists today (user just looking for a background)

I listen to rando playlists to find new music. When I show new music to my gf she does not seem to commit to liking it until she gets meta info on the artist. Its the reason I have little fear of AI music. Most listen to identity music. Politics, message, image matter to people.