@NandoDF
I seek to understand intelligence & agency and build AI aligned with compassion, freedom & universal human empowerment through progress in science & engineering
This is a fantastic presentation by @CompleteSkeptic — highly recommend it. He is spot on.
Diogo Almeida@CompleteSkeptic·We should be developing AI as a tool (1) to empower engineers and scientists to solve the big engineering challenges we face such as energy and the health of our planet (2) to empower more people to code in natural language and ensure the formal code behind the scenes is verifiable, robust, safe and secure, thus creating more jobs (3) to empower people throughout the world with access to general knowledge (4) to empower people with disabilities (5) to empower people suffering with health problems (6) to empower musicians and other artists with tools that enable them to have control and ownership of their data while allowing for them to express themselves in new creative ways (7) to empower historians and other social scientists with their research (8) to empower farmers, and others responsible for our precious food (9) to empower governments and companies with better decision making and planning to deliver better services and welfare … However, agents capable of dangerous tool use, eg autonomous weapons or hacking websites require legal scrutiny. The creators of such agents must be answerable to sovereign laws. Weaponisation of AI is a serious risk for humanity. We need diplomacy. None of the AI goals should ever require that we compromise on safety and human dignity.
I agree with this sentiment. We should be developing AI as a tool (1) to empower engineers and scientists to solve the big engineering challenges we face such as energy and the health of our planet, (2) to empower more people to code in natural language and ensure the formal code behind the scenes is verifiable, robust, safe and secure, thus creating more jobs, (3) to empower people throughout the world with access to general knowledge, (4) to empower people with disabilities, (5) to empower people suffering with health problems, (6) to empower musicians and other artists with tools that enable them to have control and ownership of their data while allowing for them to express themselves in new creative ways, (7) to empower historians and other social scientists with their research, (8) to empower farmers, and others responsible for our precious food, (9) to empower governments and companies with better decision making and planning to deliver better services and welfare, and more. Agents on the web require strong scrutiny. Weaponisation of AI is a huge risk for all humanity. It’s frightening. None of the noble goals listed above should require that we compromise on safety and human dignity and respect. Sovereign responsibility is imperative.
MeidasTouch@MeidasTouch·🚨 OBAMA ON AI: "If we are thinking about AI just in terms of how do we cure cancer or get better energy, you can do that without having agentic AI and having it just roaming free in the internet. The reason you are doing that is because you have to market a product that people will pay money for. That’s a misalignment between what our society needs and the commercial imperatives that these companies are facing, not because necessarily they’re trying to do bad things, but because they’ve got to justify these valuations. So, that’s one more reason why it is really important for us to have a competent government and a serious bipartisan conversation around this issue, and we have to do it fast. And I would encourage voters to pay attention to this. If somebody does not have a serious plan for how to deal with this, then they’re not meeting the moment, and you should probably look for somebody else."
Another amazing use of AI to empower people, and not leave anyone behind. deepmind.google/blog/putting-s…
Access to artificial intelligence should be a universal human right.
Concentration of AI in the hands of a few corporations, with a record of reckless agent experiments and militarisation, is our biggest safety risk. Sovereign AI, decentralisation, inclusion, equity, is not only the most compassionate route, it is also the safest for our children. AI empowers people. It empowers the farmers throughout the world who produce our food. It empowers girls throughout the world with less access to education because of gender bias and security. It empowers people to communicate better and deal with conflict in more productive ways. It empowers scientists, lawyers, architects, developers, doctors, nurses, ….. intelligence empowers everyone. With power there is risk, but decentralisation is the best approach we know for managing that risk.
@desirivanova Plus a long overdue consultation on non-competes and garden leaves, which the previous government refused to embrace.
I don’t like to get into politics, but I really like what the UK labour party is doing for responsible AI development in the UK. The conservatives failed. The focus is on improving security, generating jobs, and generating taxable revenue to improve our social systems. The UK has always been a leading power in science and engineering and it’s wonderful if it rises to the moment
BEING RESPONSIBLE 1. Design AI seriously for global safety and security. 2. Make safe AI accessible so everyone can benefit. Design for decentralised AI because it is more robust than the current concentration of power on the hands of a few individuals. 3. Seek international agreements to control AI for defence. I personally am against the current weaponisation of AI. It is dangerous. 4. Incentivise AI efforts that advance science, engineering, clean energy, decarbonisation, healthcare, medicine, safety and security. @periodiclabs @cusp @IsomorphicLabs etc are great examples. There is a huge cost associated with preventing AI tools to help solve some of the greatest problems we face. 5. Let’s not use AI risk hype as pre IPO marketing. It is shameful, hypocritical and irresponsible. We scientists see it. 6. Read Terence Tao’s comments on AI for mathematics - focus on building tools to enhance our understanding.
Others have expressed it well: “If you don’t have a seat at the AI table, you’re on the menu” “If you don’t own the AI model, someone else’s AI model owns you” “There are a lot of things wrong with this world… but too much intelligence is not one of them” Let’s act responsibly: design openly and seriously for global safety and security, focus AI efforts on advancing science, engineering, clean energy, decarbonisation, healthcare and security, and lastly let’s not use AI risk hype as pre IPO marketing.
James Wise@Jameswise·It's good that MPs want to discuss long tail risks, and engage with what is certainly a technology that presents both a huge opportunity and challenge for Britain. But it is wholly irresponsible to say on the one hand 'we think this is really important' and on the other block the UK's ability to build AI capabilities here and support British AI companies. In the event that neither our allies nor foes listen to the proposed international agreement and that in fact a 'kill switch' is impossible, every single one of these MPs should be fighting to strengthen Britain's hand in this critical technology by supporting local datacenters, so we're not exporting our data abroad, reforming regulations so that AI models like cyber defenses can be trained here, so we're not entirely reliant on foreign technology and supporting British AI companies, so we see the economic benefits here too. We're doing our small bit at SovAI, but this should be a national level debate, and I hope this MP champions the UK scaling those ambitions just as passionately.
If you don’t have a seat at the table, you’re on the menu. If you don’t own the model, someone else’s model owns you. I think @MarkJCarney and @j_foerst have said it well. I agree with @Jameswise on this.
James Wise@Jameswise·It's good that MPs want to discuss long tail risks, and engage with what is certainly a technology that presents both a huge opportunity and challenge for Britain. But it is wholly irresponsible to say on the one hand 'we think this is really important' and on the other block the UK's ability to build AI capabilities here and support British AI companies. In the event that neither our allies nor foes listen to the proposed international agreement and that in fact a 'kill switch' is impossible, every single one of these MPs should be fighting to strengthen Britain's hand in this critical technology by supporting local datacenters, so we're not exporting our data abroad, reforming regulations so that AI models like cyber defenses can be trained here, so we're not entirely reliant on foreign technology and supporting British AI companies, so we see the economic benefits here too. We're doing our small bit at SovAI, but this should be a national level debate, and I hope this MP champions the UK scaling those ambitions just as passionately.
Amen
Richard Sutton@RichardSSutton·There are a lot of things wrong with this world… but too much intelligence is not one of them.
This is a very important message for all enterprises in the world
Jakob Foerster@j_foerst·if you don't own your model, the model owns you. x.com/suchenzang/sta…
ENOUGH PESSIMISM IN AI PLEASE I feel that we have become unreasonably pessimistic in our field. 1. I keep hearing AI engineers saying we have to make money quickly because there’s only like 2 years left before we’re automated. Depressing. 2. I see a constant obsession with “having a moat”. This is an incredibly sad mental frame. 3. I keep hearing “we have to catch up”. Soulless. And so on. People: Every solution creates the possibility to attack new real problems. We face gargantuan engineering challenges in our world. How to capture carbon? How to get rid of teflon and plastics in water? How to invent batteries that are at least 30 times more efficient? How to solve clean energy? Better solar cells? Better ways of producing clean energy so we stop wars and famine? How to eradicate hundreds of diseases? Cures for addiction? And so on. Real engineering is about being brave and truly attacking the many problems we face, to engage with a true desire to improve the lives of others and our environment. Good engineering is not about protecting your product to make money at the expense of progress (moat thinking). Good engineering is about ensuring your children and grandchildren will be proud of the choices you made in 30 or 50 years. It is about empowering others. It is about advancing science. It is about being one step ahead. Always, one step ahead, meaningfully, proudly. These are great times. Let’s start thinking positively about all the wonderful things we could achieve together.
This is the wisest and most accurate take on scaling laws — their promise and their costly failures — that you could read today. This is what building AI systems is currently all about. Well said @jietang
jietang@jietang·Thoughts About Scaling Law Scaling, but not only of parameters. Every model release now ends with the same question: how many parameters? It isn't a question that can be answered on its own. Parameter count is only meaningful alongside three others — how much data you have, where you intend to spend your compute, and who will run the model, under what conditions. The field learned this the hard way. Kaplan et al. (2020) fit an exponent that told everyone to grow parameters faster than data — roughly 2.7:1 — and the industry complied: GPT-3, Gopher, MT-NLG. Hoffmann et al. (2022) redid the experiment across four hundred models and found the compute-optimal split is closer to 20 tokens per parameter, and that with sufficient compute the two should grow at the same rate rather than drifting apart. The error in the earlier fit compounded with every order of magnitude of compute, which is why the largest models of that generation were the most misallocated. The trillion-parameter round was, in retrospect, a detour the whole field took together and then reversed. Chinchilla wasn't the end either. It optimized training compute for models that would be trained once and evaluated. Today a model is called billions of times a day and inference dominates lifetime cost. Put inference into the objective and the optimum moves toward smaller models trained far longer — deliberate over-training, which is what Llama-2-7B and Gemma-2-9B were doing at roughly 290 and 889 tokens per parameter. Sparsity moved the target again. In a MoE model two quantities have to be kept apart: total parameters govern roughly how much the model can hold — knowledge, facts, the long tail — while activated parameters and effective depth govern roughly how far it can think, how many steps of a causal chain it can carry before it comes apart. A dense 20:1 ratio does not transfer. And the ratio isn't a single number at all: Roberts et al. (2025) find the optimal tokens-per-parameter is task-dependent, with memorization favoring more parameters and reasoning favoring more data. Follow-up work on MoE observes that at fixed TPP, pushing total parameters higher actually degrades reasoning, while activating more experts reliably helps it. This matters for what we are building toward. Finding a vulnerability is not a retrieval problem. It doesn't come from having memorized more CVEs; it comes from carrying a twenty-step chain of inference to the end without losing the thread. That capability does not live in total parameter count. Which brings us to this release. Total parameters appear to matter up to a threshold — enough to hold the world — after which additional capability comes from scaling elsewhere: effective depth per forward pass, and above all post-training. GLM-5.3 is our controlled experiment on that claim. Same base, same architecture, same total and activated parameters as GLM-5.2. One month of scaling long-horizon environments and RL. The gains are not marginal. Well, scaling has more than one dial. We turned the post-training one this time because it had the most slack left in it — not because the others are finished. Base model size, pretraining data, compute spent per forward pass: all of them are still on the table, and we will come back to each. What this experiment taught us is that the dials do not have to be turned together, and that the one worth turning next is rarely the one that was worth turning last. We are not done scaling. Next time, maybe mid-training, pre-training, and even more.
I’m super impressed with the fast progress of @inherent_labs. Worth checking out @Orbital_Ind @cusp_ai @wellingmax @ac_edwards_1
Inherent@inherent_labs·1/ Today, we introduce Faraday, a 27B-parameter AI Scientist that extends the capabilities of coding agents with a layer of scientific intuition. Trained via long-horizon RL, Faraday outperforms Claude Opus 4.8 and GPT-5.5 on the task of replicating research papers. 🧵
@jamesrichards2 @KanishkaNarayan Instead of arguing here however, how about we meet in person with other founders and VCs and share our views.
What UK AI startups need: @KanishkaNarayan In California, one can start a company the day after leaving Google. My colleagues Jeff Dean, Oriol Vinyals et al made this very clear recently with their impressive speed. In the UK, the same American companies impose 1 year garden leaves on senior AI researchers and 6 months on junior researchers. @GoogleDeepMind for example forced people to sign these contracts at the time of promotion, not the time of hiring. What this means is that researchers cannot start new companies for up to 1 year, cannot easily hire, and in short: they cannot compete. American VCs cannot understand why we move so slow. It is time for UK Gov to do the right thing: Make garden leaves optional for employees. If you’re an AI entrepreneur or VC in the UK or Europe, I would appreciate your comments here. You’ve all confided in me. It is time you let government know the urgency of this — everyone should speak up.
BTW @KanishkaNarayan — you have all our support in bringing this important change into effect. It’s an easy step, but it will transform entrepreneurship in the UK.
To be clear, @GoogleDeepMind enforces these because UK gov currently allows it. However, I think this is terrible for them too. @GoogleDeepMind has been one of the greatest places to do AI research in the UK, full of exceptional people, and a great work life balance - I highly recommend it. However, embracing these noncompete measures creates bureaucracy within the organisation, inefficiency, power trips, and altogether I don’t think this is great for anyone at @GoogleDeepMind. They should focus on doing what they can do well and best: compete.
Excessive weight decay #patagonia
Did anything happen in the world of AI while I was away with the family?
It’s my last day at Microsoft AI. It was a privilege to be part of such a talented team that achieved so much in such a short time. I’m super grateful to the @bing and @Azure teams too. @satyanadella impressed me a lot: A leader that taught us to care not only about Sota, but about delivering value to communities. I believe Microsoft AI is now in a strong place and have the talent and resources to deliver value for their users in many years to come. Thank you, so long, and looking forward to my next AI adventure too, after a precious break with my family in the Argentinian Andes.
Enhorabuena 🇪🇸 El mejor equipo ganó la copa ⚽️ 👏👏👏👏
I know everyone knows this, but what a beautiful visualisation x.com/themathflow/st…
Great point 3 by @RichardSocher —likely many bullshit jobs en.wikipedia.org/wiki/Bullshit_… won’t be impacted by AI either 🤔 x.com/richardsocher/…
What a game 🇦🇷🇦🇷🇦🇷🇦🇷🇦🇷🇦🇷🇦🇷
I loved the analysis in the CANDI (text diffusion) paper. Wonderful to see new works in this direction. However, beyond perplexity, we urgently need to start seeing benchmark results after post-training, and comparisons to autoregressive LLMs. We must measure the performance gap to know whether we’re making progress.
Any predictions for when open source GLM or DeepSeek will match Claude Fable 5? x.com/anthropicai/st…
A model trained for next-token prediction is forced to build compressed representations of latent structure in text. Ilya Sutskever correctly refers to this phenomenon as understanding. Here, a model trained for next-step sensor prediction, with a robot that has proprioception and touch sensors but no vision, is forced to build compressed representations of latent structure in the physical world. The robot becomes aware of the shape of external objects. That is, it understands the physical properties of the external world that enable it to make better next-step sensor predictions. This research was previously done by a diverse team of expert engineers at DeepMind over a month - including stars like @notmisha and @yuvaltassa. Remarkably, this reproduction with a completely different robot took only a few hours to implement using Codex. The automatic creation of physical environments by AI will likely lead to huge advances in areas of science and engineering that use physical simulators or twin models. The paper and notebook are available at ❤️∀ t.co/z2N29JF1FY
The original 2017/18 experiments by @notmisha @brandondamos @serkancabi @sergomezcol @laurent_dinh @yuvaltassa and a few other colleagues. x.com/nandodf/status…
