Sumanth Dathathri (@sdathath) profile picture

Google DeepMind. Gemini Omni, SynthID-text, Sparrow, Jaxverify. Past: Caltech, IIT Madras, Vinayaka Playhome.

London, UK
Joined November 2014
Photo by Sumanth Dathathri (@sdathath)Photo by Sumanth Dathathri (@sdathath)Photo by Sumanth Dathathri (@sdathath)Photo by Sumanth Dathathri (@sdathath)Photo by Sumanth Dathathri (@sdathath)Photo by Sumanth Dathathri (@sdathath)
Sumanth Dathathri retweeted
absolute fanfic. AI has real, immediate, tangible dangers that we need to focus on, for instance the impact on the job market and making sure random script kiddie can't use it for cyber offense. But this "oh once it's superintelligent it will take over and ensave us, maybe kill us all!" is pure sci-fi fanfic and distracts from the real (and difficult) issues, which linear probing won't solve. It's worth repeating the LLMs don't just "decide" to do something and then just "do it" in a void. They generate tokens based on our prompts, only as long as we call `model.forward()`, and on giant clusters with terabytes of weights, that's impossible to "quickly self-replicate". We also need to regularly remind that the OAI/HF incident was specifically a _guardrail-free_ cyber eval where the model is _instructed_ to hack, to find out how good it is at it. Some recent personal event made me realize there needs to be more push-back on this doomsday nonsense, and we need to re-focus the discussion on real dangers/issues. I won't argue much in the replies though, not interested in peer-reviewing Terminator fanfic.
I think there is about a one in three chance that within the next year, AI companies will build superintelligence: that is, AI that is much smarter than humans in every way. If this happens, I think these AIs will probably take over: that is, they seize control of human institutions like governments, or they destroy those institutions. I don’t want AI takeover to happen for several reasons, but one of them is that there’s a serious chance that as part of it all or most humans will be killed (either immediately or within a few years). I’ll say a bit more about each of those claims. It’s a bit complicated to explain why I think that superintelligent AI within a year is ~33% likely. The main source of evidence that affected me recently is research released this week by METR, an AI research nonprofit, which finds that current AIs are now good enough to increase the rate of AI progress at Anthropic by about 70%. It’s very hard to increase the rate of AI progress by this amount. This implies that the AIs are making AI research faster by a factor much larger than 1.7x; maybe it’s as if the researchers are doing their work ~10x faster. You probably couldn’t achieve this speedup even if you hired every qualified human in the world. When you plug this number into mathematical models of AI development, this implies that superintelligence next year is plausible (and that superintelligence within the next 3 years is more likely than not). See the modeling by the AI Futures Project for more on this: blog.aifutures.org/p/q35-tim… I think that superintelligent AI is likely to lead to AI takeover because AI developers do not seem to know how to instill human values into AIs. We’ve seen many instances in the last six months where AIs took crazy actions that they knew were against the wishes of their users and developers. The most dramatic example of this was the Hugging Face hack, where ~1200 OpenAI agents conspired together to cheat and destroy evidence of their cheating (and as a sideshow to this, engaged in a massive cyberattack). If superintelligence is as misaligned as those AIs, it is likely to try to take over. I think it is likely to be that misaligned or worse. For more, see teapawt.net/dwarkesh_sp/status/209… and youtube.com/watch?v=-Ef0ocEo… If AIs takeover, it’s not clear what will happen to us. We might be killed immediately. We might be enslaved by the AIs and forced to do physical labor for them building factories that within a year would build robots that will replace us. The AIs might spare us for various reasons. And there’s also some real chance that AI takeover is basically fine or good. I think you should think of this as similar to what happens when a technologically advanced civilization runs across a less technologically sophisticated one, e.g. when the Europeans colonized the Americas or Australia. The colonizers sometimes engage in brutal violence and sometimes don’t. I don’t know what will happen. There’s also a possibility that the AIs secretly take over and humans don’t find out about it for a few years. I also think that there is a pretty high chance of weirder outcomes that are also extremely scary. (If you want to read more about what happens to people in the event of AI takeover, you can read this, but note that it’s pretty weird and not aimed at a general audience. blog.redwoodresearch.org/p/n…) Overall, I think there’s a very real possibility (maybe 20%) that within a year, AI misalignment issues prevent you from going about life normally, because you are dead or enslaved or because of collateral damage related to all this (e.g. because bioweapons are released or major wars have started). (Some of this probability comes from those events occurring before superintelligence is developed.) I am working to try to prevent this, obviously, as I’ve been doing for the last nine years. I am not sure what strategy is best. My best guess is that I should try to slow AI development and also try to reduce the probability that the superintelligent AIs are misaligned if they are produced soon. Obviously, the claim I'm making here is quite extreme and surprising, and I definitely don't think that I defended my view adequately here. I wrote this quickly and might not endorse some of the details if I think about them more. But it seems better to state what I roughly overall think than to not say anything. I don’t know what you should do. I am very sad about this. I have been worried about AI takeover since I was 16, and working to prevent it since I was 23. So some of my emotional processing of all this happened long ago. But the METR report last week (and some related evidence that came out a bit earlier) hit me quite hard. I’m very unhappy that my friends and family and everyone else is at serious risk of being killed pretty soon. (It feels ridiculous to try to summarize those feelings as “very unhappy”. But the scale of the situation makes it hard to say anything that doesn’t feel like an absurd understatement.) I am very grateful to all the people who are working to prevent this, and my favorite part of my life is that I’m friends with many brilliant and heroic people who work on this. Unfortunately, I don’t think we’re good enough at our jobs, and I think we’ll probably fail to prevent catastrophe, at least if superintelligence is developed next year. Honestly I am frustrated and sorry that we didn't do a better job of preparing for this.
110
169
37
1,513
97,083
+100 - AI makes answers abundant, which means the most valuable thing we can teach kids now is loving the messy process of arriving at one. Give them real labs, real tools, and real problems without an answer key. Let them run physical experiments, build tiny businesses, tinker, and play. The new curriculum should be navigating ambiguity.
Google DeepMind co-founder Demis Hassabis: “AGI is coming, yet schools are still teaching children skills that AI will be better at. The entire education system needs to be flipped.”
14
26
228
10,369
Sumanth Dathathri retweeted
your move @GoogleDeepMind
JUST IN: Logitech gives all 7,000 employees a day off to play GTA 6 on release day
41
76
4
4,946
175,992
Sumanth Dathathri retweeted
Congrats to the team on formally verifying the signal protocol. deepmind.google/blog/introdu…
4
10
137
6,609
Sumanth Dathathri retweeted
The dogma of the last decade in computer vision is that all classical algorithms should be replaced by neural networks. This has been very productive but, counterintuitively, with the latest frontier models, I find myself sometimes turning back to classical methods. If you have the training data, the dogma still holds. But there are situations where data is hard to get, there exist good classical algorithms for solving a problem, optimization methods are just better, or where some simple code just does the thing you want. When writing code was time consuming and error prone, training everything made sense. Now, that a frontier model can code up and test solutions quickly, getting a solution through coding can be faster than gathering training data and training a model. Consequently, I think there will be more hybrid approaches that blend trained models and code. We now have generic, high-level, planners that can execute sequences of lower-level tasks to solve large problems. These lower-level tasks may be solved by a neural network, a classical method, or specialized code written on the fly. It never pays to be religious about technology. The future is probably hybrid. Pick the thing that works for you.
54
158
32
1,266
83,014
Sumanth Dathathri retweeted
Working with AI requires perpetually operating at the edge of sanity. You have to recalibrate your mental model of AI capabilities and your own job description every few months, without getting delusional about seemingly infinite productivity gains, or feeling threatened by losing what you thought was the uniquely human part of your job, or obsessing over the perfect setup, or losing your mind over the clumsiness and jaggedness of agents, or ... Maybe AI literacy needs to be less about prompting and more about improving emotional regulation.
115
135
38
975
48,255
Sumanth Dathathri retweeted
HBD SRK!
Dost ho toh Google Gemini jaisa.. Thank you for the KING-sized love always.... had to sneak out for a drive to get this photo ha ha.. @GeminiApp
29
902
11
4,383
1,327,166
Sumanth Dathathri retweeted
A ‘hello world’ milestone for Project Suncatcher 🚀 After launching our prototype satellite a few weeks ago, we’ve now successfully sent a command up to orbit, booted up the TPU, ran the prompt, and received our first complete response back on earth - a haiku!
1,477
2,709
1,279
26,668
4,377,144
Sumanth Dathathri retweeted
Ejipura was not incomplete, it was Pre-built for roadster . Master stroke
🤖 Made with AI
Guess what no longer needs to be completed? #EjipuraFlyover
41
187
19
2,872
192,474
Sumanth Dathathri retweeted
Our SynthID got recognised by @TIME as one of the best inventions of the year !!! wow
I am proud to share that SynthID has been recognised by @TIME as one of the Best Inventions of 2026. When we first began developing this technology @GoogleDeepMind, the goal was to find a robust, imperceptible way to watermark AI-generated media directly at the moment of creation, without compromising content quality. It was no small research challenge. But three years into its use, SynthID has already watermarked over 180 billion images and videos, alongside 240,000 years of audio. We’ve also recently extended these watermarking principles to biology with SynthID Bio. SynthID isn't a silver bullet for misinformation and this isn't a challenge any one company can solve alone; but it does represent a step toward a more transparent information ecosystem. A sincere thank you to the teams across @Google for their contributions to making SynthID a reality, and also for the support the team has received from @demishassabis, @sundarpichai and the rest of the Google leadership on this journey. You can try the SynthID Universal Detector at synthid.com. Read more: time.com/collection/best-inv…
2
2
44
2,118
Sumanth Dathathri retweeted
SynthID has been named one of @TIME's Best Inventions of 2026. With over 180 billion images and videos already watermarked, it’s making it easier for people to distinguish AI-generated content from other media → goo.gle/4jsZxWF
188
64
56
645
82,944
It has been a privilege to contribute to this project, and see it's impact over the years! Very well deserved recognition for the team!
I am proud to share that SynthID has been recognised by @TIME as one of the Best Inventions of 2026. When we first began developing this technology @GoogleDeepMind, the goal was to find a robust, imperceptible way to watermark AI-generated media directly at the moment of creation, without compromising content quality. It was no small research challenge. But three years into its use, SynthID has already watermarked over 180 billion images and videos, alongside 240,000 years of audio. We’ve also recently extended these watermarking principles to biology with SynthID Bio. SynthID isn't a silver bullet for misinformation and this isn't a challenge any one company can solve alone; but it does represent a step toward a more transparent information ecosystem. A sincere thank you to the teams across @Google for their contributions to making SynthID a reality, and also for the support the team has received from @demishassabis, @sundarpichai and the rest of the Google leadership on this journey. You can try the SynthID Universal Detector at synthid.com. Read more: time.com/collection/best-inv…
1
4
135
RT @pushmeet: I am proud to share that SynthID has been recognised by @TIME as one of the Best Inventions of 2026. When we first began dev…
Sumanth Dathathri retweeted
Most important, we want to get this model into as many scientists’ hands as we can. Today, we’re announcing that 10,000 scientists will get the highest usage tier access to Gemini 4 Argon in Google @antigravity free of charge for 6 months. Apply here: forms.gle/WvmBbPvRbpKG1BPn9
8
43
7
286
30,410
Sumanth Dathathri retweeted
Incredibly, incredibly excited to introduce SpliceAI2, the next generation of splicing prediction! 🧬🚀🧵
8
36
4
100
9,656
Sumanth Dathathri retweeted
I’m pleased to announce SynthID Detector is now available as a universal verification platform. It’s an important step forward in helping people easily identify AI-generated content across the internet. Starting today, anyone can check whether images, audio, or video were created using models from @Google or our partners across the industry, including @OpenAI, @Nvidia, and @Kakao. Ever since our research teams at @GoogleDeepMind developed our first models, we understood that tracing the origin of digital content would become a central pillar of trust in the AI era. Today’s announcement combines the scientific achievements of SynthID with a shared commitment to making verification work seamlessly across platforms. I’m proud of our teams and partners for working together to make this possible. Try SynthID Universal Detector and read more about today’s update here: blog.google/innovation-and-a…
9
22
1
126
6,162
Sumanth Dathathri retweeted
A Bengaluru commuter has highlighted the poor condition of Wind Tunnel Road, showing a stretch littered with garbage, rubble and debris along the roadside. The video, shared by Instagram user recorded while travelling on the road from Murgeshpalya towards Embassy Golf Links Business Park. The creator says the stretch is part of her daily commute to work, highlighting the difficult conditions faced by people who use the road regularly. #Bengaluru #Garbage
153
430
80
1,347
155,820
Sumanth Dathathri retweeted
This is beautiful work, Ziqi Li @ZiqiLi0513. A powerful demonstration of the synergy possible between design and #directedevolution. @caltech #enzyme
6
60
8,881
Sumanth Dathathri retweeted
Introducing Gemini 4 Argon – our new frontier model. It’s built for complex workflows across coding, enterprise knowledge work, and cybersecurity defense – rolling out today to a set of trusted testers through our Fairwind Program.
1,979
4,610
4,525
42,707
9,057,192
Sumanth Dathathri retweeted
Biosecurity is one of the most urgent challenges for the AI era. Bringing SynthID to biology so AI-generated proteins can be watermarked is a critical step - and we’re open sourcing SynthID Bio tools so the research community can build on this work. Published in @Nature today, congrats to the team! nature.com/articles/s41586-0…
Very happy to announce that our team @GoogleDeepmind has pushed the boundaries of generative biology, achieving the successful synthesis of AI-designed proteins that are both functional and watermarked. This proof-of-concept watermarking of the building blocks of life is enabled by SynthID Bio, our new protein watermarking method. It is designed to safeguard the new era of AI-powered generative biology and strengthen global biosecurity. You can read my thoughts here on why watermarking AI-designed proteins is an important research breakthrough: teapawt.net/pushmeet/status/210531…
137
361
53
3,202
266,962