The Thinking Game Film – Google DeepMind documentary
(thinkinggamefilm.com)210 points by ChrisArchitect 3 days ago
210 points by ChrisArchitect 3 days ago
The chatbots and image editors are just a side-show. The real value is coming in e.g. chemistry (Alpha fold etc all), fusion research, weather prediction etc.
None of that has reached the market yet. If it was up to the sciences alone, AI couldn't bear the weight of its own costs.
It also needs to be vertically integrated to make money, otherwise it's a handout to the materials science company. I can't see any of the AI companies stretching themselves that thin. So they give it away for goodwill or good PR.
Science in general tends to be subsidised and given away because as basic understanding of the world is hard to monopolise. I'm not sure how Einstein would have done a general relativity startup.
That said Deepmind are doing a spin-off making drugs https://www.isomorphiclabs.com/
That's not really true. Commercial weather prediction has reached the market, and a drug (sorry, can't find the new s link) that was found by AI-accelerated drug discovery is now in clinical testing
The reason why vertical integration is important for AI investment is that if AI is commoditized, then that AI-acceleration will costs pennies for drugs that are worth billions.
I don't see how OpenAI or Google can profit from drug discovery. It's nearly pure consumer surplus (where the drug companies and patients are the consumers).
Right. More accurate predictions for meta-data based killings which as championed by US in their war on terror
LLMs in general are ML based, need a lot of data and compute. The same infrastructure as any other ML based system.
The AI/AGI hype in my opinion could be better renamed to ml with data and compute 'hype' (i don't like the word hype as it doesn't fit very well)
unfortunately all this work on sora has very real military use case. I personally think all this investment in sora by open AI is largely to create a digital fog of war. Now when a rocket splatters a 6 year old palestinian girl's head across the pavement like a jackson polock painting, They will be able to claim its AI generated by state sponsored actors in order to prevent disruption to the manufactured consent aperatus.
Why are images and video a complete waste? This makes no sense to me.
Right now the generators aren’t effective but they are definitely stepping stones to something better in the future.
If that future thing produces video, movies and pictures better than anything humanity can produce at a rate faster than we can produce things… how is that a waste?
It can arguably be bad for society but definitely not a waste.
Let me phrase it a bit differently, then: AI generated cats in Ghibli style are a waste, we should definitely do less of that. I did not hold that opinion before the documentary
Education-style infographics and videos are OK.
I’m not even talking about this. Those cat videos are just stepping stones for academy award winning masterpieces of cinema like dune. All generated by AI on a click in one second.
It might be shocking to you but some people believe there is more to life than producing and consuming "content" faster and faster.
Most of it is used to fool people for engagement, scam, politics or propaganda, it definitely is a huge waste of resource, time, brain and compute power. You have to be completely brainwashed by consumerism and techsolutionism to not see it
I see it. But you’re lacking imagination to what I’m referring to. It’s also fucking obvious. Like I’m obviously not referring to TikTok videos and ads and that kind of bullshit every one on earth knows about and obviously hates. You’re going on as if it’s “shocking” to me when what you’re talking about is obvious as night and day. What’s shocking to me is that you’re not getting my point and I’m obviously talking about something less well known.
Take your favorite works of art, music and cinema. Imagine if content on that level can be generated by AI in seconds. I wouldn’t classify that as a “waste” at all. You’re obviously referring to bullshit content, I’m referring to content that is meaningful to you and most people. That is where the trendline is pointing. And my point, again is this:
We don’t know the consequence of such a future. But I wouldn’t call such content created by AI a waste if it is objectively superior to content created by humans.
I actually had a counter thought a few years ago.
We consume A LOT of entertainment every day. Our brains like that a lot.
Doesn't has to be just video but even normal people not watching tv at all entertain themselves through books or events etc.
Live would be quite boring.
Parent said "entertainment use cases" are a complete waste, not all uses of images and video. I don't agree, but do particularly find educational use cases of AI video are becoming compelling.
I help people turn wire rolling shelf racks into the base of their home studio, and AI can now create a "how to attach something to a wire shelf rack" without me having to do all the space and rack and equipment and lighting and video setup, and just use a prompt. It's not close to perfect yet, but it's becoming useful.
> particularly find educational use cases of AI video are becoming compelling.
compelling graphics take a long time to create. for education content creators, this can be too expensive as well. my high school physics teacher would hand draw figures on transparencies on an overhead projector. if he could have produced his drawings as animations cheap and fast using AI, it would have really brought his teaching style (he really tried to make it humorous) to another level. I think it would be effective for his audience.
imagine the stylized animations for things like the rebooted Cosmos, NOVA, or even 3Blue1Brown on YT. there is potential for small teams to punch above their weight class with genAI graphics
If AI can produce movies, video and art better aka “more entertaining” then humans than how is it a waste?
DeepMind's new [edit: apparently now old] weather forecast model is similar in architecture to the toys that generate videos of horses addressing Congress or cats wearing sombreros. The technology moves forward and while some of the new applications are not important, other applications of the same technology may be important.
Is it really similar? I was under the impression it's a GNN of a (really dense) polyhedron, not a diffusion model
GenCast is a diffusion model, but it is not the "new" one like I said. Apparently there is another one. https://arxiv.org/pdf/2506.10772
reposting this from youtube comment
From 1:14:55-1:15:20, within the span of 25 seconds, the way Demis spoke about releasing all known sequences without a shred of doubt was so amazing to see. There wasn't a single second where he worried about the business side of it (profits, earnings, shareholders, investors) —he just knew it had to be open source for the betterment of the world. Gave me goosebumps. I watched that on repeat for more than 10 times.
Another way to interpret this (and I don't mean it pejoratively at all): Demis has been optimizing his chances for winning a nobel prize for quite some time now. Releasing the data increased that chance. He also would have been fairly certain that the commercial value of the predictions was fairly low (simply predicting structures accurately was never the rate-limiting step for downstream things like drug discovery). And that he and his team would have a commercial advantage by developing better proprietary models using them to make discoveries.
My interpretation of that moment was that they had already decided to give away protein sequences as charity, it was just a decision of all as a bundle vs fielding individual requests (a 'service').
Still great of them to do, and as can be seen it's worth it as a marketing move.
(as an aside, this is a common thing that comes up when you have a good model: do you make a server that allows people to do one-off or small-scale predictions, or do you take a whole query set and run it in batch and save the results in a database; this comes up a lot)
DB of known proteins is not where the money can be made, designing new proteins is. This is why AlphaFold3 (that can aid in this) is now wrapped in layers of legalese preventing you to actually use it in the way you want. At least that's what my lifescience users tell me. Big Pharma is now paying Big Money to DeepMind to make use of AF3 ...
I also noticed this as well. Actually went back and watched it several times. It's an incredible moment. I keep thinking, "if this moment is real, this is truly a special person."
Greg Kohs and his team are brilliant. For example, the way it captured the emotional triumph of the AlphaFold achievement. And a lot of other things.
One of the smart choices was that it omitted a whole potential discussion about LLMs (VLMs) etc. and the fact that that part of the AI revolution was not invented in that group, and just showed them using/testing it.
One takeaway could be that you could be one of the world's most renowned AI geniuses and not invent the biggest breakthrough (like transformers). But also somewhat interesting is that even though he had been thinking about this for most of his life, the key technology (transformer-type architecture) was not invented until 2017. And they picked it up and adapted it within 3 years of it being invented.
Also I am wondering if John Jumper and/or other members of the should get a little bit more credit for adapting transformers into Evoformer.
Is that a good thing or a bad thing? Demis is after all a co-founder and CEO.
Makes it seem that AI is a one-man show while also feeding the hype cycle
Our society is leader based. Otherwise the garbage Trump wouldn't matter but he does. The same thing with garbage Musk. Musk gets what he wants from Tesla because the shareholders believe that Musk is critical to Tesla.
Both are fundamental to their followers.
So its quite clear that you can't just say 'its DeepMind' but have a figure in the middle of it like Dennis.
They trust him to lead DeepMind.
Watched it this week. Pretty good.
There are a couple parts at the start and the end where a lady points her phone camera at stuff and asks an AI about what it sees. Must have been mind-blowing stuff when this section was recorded (2023), but now it's just the bare minimum people expect of their phones.
Crazy times we're living in.
I was ok with that as "fledgling AI" at the start of the movie/documentary, but thought that going back to it and having the chatbot suggest a chess book opening to Hassabis at the end was cheesy and misleading.
They should have ended the movie on the success of AlphaFold.
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5409063
AI for science is much bigger than RL or Generative AI in science.
There are several classes of models Like operator learning, physics informed neural networks, Fourier operators
That perform magnificently well and have killer applications in various industrial settings
Do read the attached paper if you're curious about AI in science
I caught it on the airplane a few days ago. I would have loved a little more technical depth, but I guess that's pretty much standard for a puff piece.
It is interesting that Hassabis has had the same goal for almost 20 years now. He has a decent chance of hitting it too.
It's official too. It's on: https://www.youtube.com/@googledeepmind
Moderators: Please change the link; feels kind of unethical to bait someone into paying for this now.
Streaming on YouTube now: https://www.youtube.com/watch?v=d95J8yzvjbQ
What confused me about this documentary was the "at home" scenes for Hassabis.
He is famously a North London lad, but the at home shots are clearly shot from South London looking North (you can tell by the orientation of The Shard and Bishops Gate out of the window).
I thought that this might have been a "stage home" but it appears to be the same place in the background of various video conferences he is on too, so unless those were staged for the documentary (which seems like a lot of effort), then he lives near Crystal Palace and not Highgate?
I find it funny that the YouTube link takes you to the film, but like an hour into it.
Is the multimodal agent really as good as shown in the documentary? If so, why did Google need to stage parts of the demo at Google I/O?
Earlier on HN: https://news.ycombinator.com/item?id=46086561
Hard to discount the impact of AlphaFold in science work but submitting this to a number of film festivals like Tribeca seems a bit AI-washing.
There's some funny comments going on in this thread. Understandably so. What could be more divisive an issue than AI on a silicon valley forum!?
As a brit, I found it to be a really great documentary about the fact that you can be idealistic and still make it. There are, for sure, numerous reasons to give Deepmind shit: Alphabet, potential arms usage, "we're doing research, we're not responsible". The Oppenheimer aspect is not to be lost, we all have to take responsibility for wielding technology.
I was more anti-Deepmind than pro before this, but the truth is as I get older it's nicer to see someone embodying the aspiration of wanton benevolence (for whatever reason) based on scientific reasoning, than to not. To keep it away from the US and acknowledge the benefits of spreading the proverbial "love" to the benefit of all (US included) shows a level of consideration that should not be under-acknowledged.
I like this documentary. Does AGI and the search for it scare me? Hell yes. So do killer mutant spiders descending on earth post nuclear holocaust. It's all about probabilities. To be honest: disease X freaks me out more than a superintelligence built by an organisation willing to donate the research to solve the problems of disease X. Google are assbiscuits, but Deepmind point in the right direction (I know more about their weather and climate forecasting efforts). This at least gave me reason to think some heart is involved...
i tried to watch it but like AI in general, it was extraordinarily boring. neural nets are really cool technically, but the whole AI thing is just getting old and I couldnt care less where its going
we can guarantee that whether its the birth of superintelligence or just a very powerful but fundamentally limited algorithm, it will not be used for the betterment of mankind, it will be exploited by the few at the top at the expense of the masses
because thats apparently who we are as a species
Hi, I’m genuinely curious about your writing style. I’m seeing this trend of no proper casing and no punctuation becoming vogue-ish. Is there a particular reason you prefer to write this way or is this writing style typical for a generation? Sincere question, not snark, coming from an older generation guy.
This is the writing style of this generation. I've just scrolled 6 months of my conversation with a friend in his twenties. Not a single comma or period to be seen. I mean on his side.
it signals high status and nonconformity. the reader intuits that a sigma male is speaking and he doesnt play by the rules. hes not bound by the constraints and regulations of classical reality. hes dangerous
but seriously, its just more comfortable to type. apostrophes and capitals are generally superfluous, we'll and well the only edge case, theyve, theyll, wont, dont etc its just not necessary. theres no ambiguity
i only recently started using full stops for breaks. for years, I was only using commas, but full stops are trending among the right people. but only for breaks, not for closing
I’m certainly no authority but i tend to write the same way for casual communication, came from the 90s era BBS days. It was (and still is) common on irc nets too. Autocorrect fixes up some of it, but sometimes i just have ideas i’m trying to dump out of my head and the shift key isn’t helping that go faster. Emails at work get more attention, but bullshittin with friends on the PC? No need.
I’ll code switch depending on the venue, on HN i mostly Serious Post so my post history might demonstrate more care for the language than somewhere i consider more causal.
If you watch on there's a bit where they decide to give away all the protein folding results for free when they could have charged (https://youtu.be/d95J8yzvjbQ?t=4497). Not everything is exploitation rather than the betterment of mankind.
that sort of mentality is typical in researchers, but the powers that be will be thinking about profit and control, mass layoffs and AI governance in conjunction with digital id, carbon credits etc
every technological advancement that made people more productive and should have led to them having to do less work, only led to people needing to do more work to survive. i just dont see AI being any different
Correct! I’m glad people are finally starting to get it
Its so disappointing to read this.
Do you know how long it took us to get to this point? Massive compute, knowledge, alogorithm etc.
Why are you even on HN if the most modern and most impactful technologie leads you to say "i couldn't care less were its going'?
Just a few years ago there was not a single way of just solving image generation, music generation and chat bots which actually able to respond reasonable to you and that in different languages.
AlphaFold already helps society today btw.
AlphaFold is optimization, not thinking. Propaganda 'r us.
Did you watch the documentary? Would probably fare better if you did, because it'd give you the context for the film title.
I'm an hour into it, unconvinced.
The illusion that agency 'emerges' from rules like games, is fundamentally absurd.
This is the foundational illusion of mechanics. It's UFOlogy not science.
Well, two things: it's the last sentence of the film; being on hour into something you're calling propaganda is brave.
Anyways. I thought the documentary was inspiring. Deepmind are the only lab that has historically prioritized science over consumer-facing product (that's changing now, however). I think their work with AlphaFold is commendable.
Agency will emerge from exceeding the bottleneck of evolution's hand-me-down tools: binary, symbols, metaphors. As long as these unconscious sportscasters for thought "explain" to us what thought "is", we are trapped. DeepMind is simply another circular hamster wheel of evolution. Just look at the status-propaganda the film heightens in order to justify the magic.
Quite honestly, it's about time the penny dropped.
Look around you, look at the absolute shit people are believing, the hope that we have any more agency than machines... to use the language of the kids, is cope.
I have never considered myself particularly intelligent, which, I feel puts me at odds with many of HN readership, but I do always try to surround myself with myself with the smartest people I can.
The amount of them that have fallen down the stupidest rabbit holes i have ever seen really makes me think: as a species, we have no agency
Not sure why this is downvoted. The comment cuts to the core of the "Intelligence vs. Curve-Fitting" debate. From my humble perspective as a PhD in the molecular biology /biophysics field you are fundamentally correct: AlphaFold is optimization (curve-fitting), not thinking. But calling it "propaganda" might be a slight oversimplification of why that optimization is useful. If you ask AlphaFold to predict a protein that violates the laws of physics (e.g. a designed sequence with impossible steric clashes), it will sometimes still confidently predict a folded structure because it is optimizing for "looking like a protein", not for "obeying physics". The "Propaganda" label likely comes from DeepMind's marketing, which uses words like "Solved"; instead, DeepMind found a way to bypass the protein folding problem.
If there's one thing I wish DeepMind did less of, it's conflating the protein folding problem with static structure prediction. The former is a grand challenge problem that remains 'unsolved' while the latter is an impressive achievment that really is optimization using a huge collection of prior knowledge. I've told John Moult, the organizer of CASP this (I used to "compete" in these things), and I think most people know he's overstating the significance of static structure prediction.
Also, solving the protein folding problem (or getting to 100% accuracy on structure prediction) would not really move the needle in terms of curing diseases. These sorts of simplifications are great if you're trying to inspire students into a field of science, but get in the way when you are actually trying to rationally allocate a research budget for drug discovery.
I'm really curious about this space: what types of simulation/prediction (if any) do you see as being the most useful?
Edit to clarify my question: What useful techniques 1. Exist and are used now, and 2. Theoretically exist but have insurmountable engineering issues?
It seems that to solve the protein folding problem in a fundamental way would require solving chemistry, yet the big lie (or false hope) of reductionism is that discovering the fundamental laws of the universe such as quantum theory doesn't in fact help that much with figuring out the laws/dynamics at higher levels of abstraction such as chemistry.
So, in the meantime (or perhaps for ever), we look for patterns rather than laws, with neural nets being one of the best tools we have available to do this.
Of course ANNs need massive amounts of data to "generalize" well, while protein folding only had a small amount available due to the months of effort needed to experimentally discover how any protein is folded, so DeepMind threw the kitchen sink at the problem, apparently using a diffusion like process in AlphaFold 3 to first determine large scale structure then refine it, and using co-evolution of proteins as another source of data to address the paucity.
So, OK, they found a way around our lack of knowledge of chemistry and managed to get an extremely useful result all the same. The movie, propaganda or not, never suggested anything different, and "at least 90% correct" was always the level at which it was understood the result would be useful, even if 100% based on having solved chemistry / molecular geometry would be better.
We have seen some suggestion that the classical molecular dynamics force fields are sufficient to predict protein folding (in the case of stable, soluble, globular proteins), in the sense that we don't need to solve chemistry but only need to know a coarse approximation of it.
I'm concerned that coders and the general public will confuse optimization with intelligence. That's the nature of propaganda, substituting sleight of hand to create a false narrative.
btw an excellent explanation, thank you.
What's the difference between optimisation and intelligence?
There is quite a bit of bait-and-switch in AI, isn't there?
"Oh, machine learning certainly is not real learning! It is a purely statistical process, but perhaps you need to take some linear algebra. Okay... Now watch this machine learn some theoretical physics!"
"Of course chain-of-thought is not analogous to real thought. Goodness me, it was a metaphor! Okay... now let's see what ChatGPT is really thinking!"
"Nobody is claiming that LLMs are provably intelligent. We are Serious Scientists. We have a responsibility. Okay... now let's prove this LLM is intelligent by having it take a Putnam exam!"
One day AI researchers will be as honest as other researchers. Until then, Demis Hassabis will continue to tell people that MuZero improves via self-play. (MuZero is not capable of play and never will be)
Sure, but AlphaFold is still probably the most impactful and positive thing to have come out of "Deep Learning" so far.
Didn’t the transformer model come from AlphaFold? I feel like we wouldn’t have had the LLMs we use today if it wasn’t for AlphaFold.
The Transformer was invented at Google, but by a different team. AFAIK the original AlphaFold didn't use a transformer, but AlphaFold 2.0 and 3.0 do.
Sharp wave ripples, nested oscillations, cohering at action-syntax. The brain is "about actions" and lacks representations.
Creatively peeling the hyper dimensional space in the scope of simplectic geometry, markhov blanket and helmholtz invariance????
Watched it a while ago. Made me seriously think about AI and what we should use it for. I feel like all the entertainment use cases (image and video gen) are a complete waste.