@_sholtodouglas
Scaling RL @AnthropicAI, ex @DeepMind - working towards intelligence too cheap to meter
people have been asking 'if models are so good where is all the new software' for the last year or so. As always, look to gaming for early evidence of the future
Peter Schmidt-Nielsen@ptrschmdtnlsn·What a bizarre future. I wanted to play Halo: Combat Evolved for the first time with @robertskmiles, and we wanted local multiplayer across two screens, and a model could just immediately add that feature to the decompilation.
the high frequency personal agent wars of 2027, fought via low latency chips and rack placement next to the servers hosting Resy and SF tennis court systems
we're witnessing a new artform and artist created live
Tobin South@TobinSouth·I'm calling it now, this will be the song of the (sf) summer x.com/anabology/stat…
Fantastic conversation - Joe asks tough questions! Did we answer them well? I think this might be the first appearance of @the_marwell in public - he is a complete legend within Anthropic, an unbelievable amount of critical programs simply would not have progressed without him. Getting to know him via this podcast is crucial lore.
Joe Lonsdale@JTLonsdale·Some of my friends will be mad I recorded this, and comms people at Anthropic objected to releasing certain parts (and delayed it). But it's important for leaders to make these conversations happen. And I have a lot of respect for these two. Over a month ago, I sat down with Anthropic's key technical leaders @_sholtodouglas & @the_marwell for an optimistic insiders’ view of the AI frontier, with hard questions too: open source, regulatory capture, slowing down USA vs China, and more.
Casey’s ‘Domes are overrated’ blog post is an all-timer caseyhandmer.wordpress.com/2019/11/28/dom…
Casey Handmer@CJHandmer·This is too much fun. A render of how the quilted hyperbolic isotensoid Mars city structure might work.
This is excellent 🫡 Its pretty fucking cool that we get to live at the time when there is a "White House Accord on Super Intelligence." Lets work together and do this right.
Hell yeah.
Pieces of the bio-shield coming together! Hannu also happens to be the author of one of my favorite sci fi series, and you should read gwern.net/review/quantum… to instantly see how brilliant he is
Hannu Rajaniemi@hannu·With @WSJ and @georgia_wells we are now sharing a bit more about our work at @redqueenbio : using AI and a lot of hard wet lab work to create medicines that let us prepare for both natural and synthetic viruses in advance. x.com/WSJ/status/210…
so much code
aidan@aidanshandle·Claude deeply understands the physical world and can simulate it to an astonishing level of detail.
aidan@aidanshandle·This whole shot is code. Claude wrote all of it, renderer and physics, from scratch x.com/gallabytes/sta…
Claude deeply understands the physical world, and can create simulations at astonishing levels of detail.
aidan@aidanshandle·how claude did it, explained by claude x.com/gallabytes/sta…
Kiana Ehsani@ehsanik·Since I am obsessed with painting these days I got Opus 5.5 to replicate Mona Lisa. This was my prompt: "can you search for the picture of Mona Lisa, then replicate it using mouse movement and brushes in JS paint using chrome?" It took 20 minutes for the drawing part using Cowork + Claude in Chrome. After it was done I asked: "Ok now can you make a video of exactly what you just did? Showing every mouse movement and changes?" Result is the video you see! :D More details in thread 🧵
We can get to a defence dominant world if we build it
Jake Adler@jakeadler·1/ Biological threats can bring a country to a halt without a single shot being fired. Yet much of our biodefense infrastructure has barely changed in decades, even as America’s institutions produce extraordinary scientific breakthroughs.
When jack showed me this I was completely blown away. Want to guess how it is made?
theseriousadult@gallabytes·feeling the agi lately
this is so funny
NotinReality@other__reality·Claude Opus 5.5 has the best visual design of any model I have tested so far x.com/slimer48484/st…
its interesting how claude manually wielding the brush leads to a very different form of creativity than diffusion based models (and it shows how deeply this generation of claude can _see_)
Jake Eaton@jkeatn·for the past few months i've been asking our models to paint. opus 5.5 is very skilled at emulating different styles every image here is a python program generated pixel by pixel. there is no image model, and no off-the-shelf art software. instead, it's about 7,500 lines of code using standard libraries to emulate different brush styles. the agents don't use any pictures as reference, instead working only from what they know about each painter
the 5.5 series have a serious step up in their ability to understand and model in 3D
Ben Poole@poolio·sketch-to-simulation with Opus 5.5 x.com/claudeai/statu…
Wes Bos@wesbos·I've been using Opus 5.5 over the last week and ran a TON of tokens through it. I can tell you it's def on the fable level It's VERY good at 3D modeling. I had it render the top 500 gym equipment pieces and create a whole gym builder x.com/claudeai/statu…
lfg - terraforming mars will be one of the big quests starting in the 2030s
Erika Alden DeBenedictis@erika_alden_d·Today, Pioneer Labs is announcing our first step towards terraforming Mars. 🚀🌼 With equipment that fits in just a single rocket launch, we can convert Martian dirt, water, and air into enough building materials to construct a small city on Mars. To do it, we made the first microbe for Mars. We found the best microbe on Earth and used evolution to teach it how to source all of its nutrients directly from Martian materials. The first astronauts will be greeted with safe shelter already filled with water, oxygen, and rocket fuel for the return journey. This is the first step toward using biology to make Mars a friendly place for life. It lets us live off the land and helps us build the next great frontier. It's the first of five organisms we need to green Mars ⬇️
also important news we fixed the writing
Claude@claudeai·Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family. It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.
Claude@claudeai·Opus 5.5 communicates more naturally, addressing some of the most common feedback we heard on Opus 5. It puts the most important information up front and follows the writing rules you give it, which makes long sessions easier to follow.
these are fun to think about, a nice mountain to climb
Sam Rodriques@SGRodriques·In light of the progress in mathematics, we at Edison Scientific and FutureHouse have assembled a set of Millennium Problems for Biology. They are chosen to be very hard to solve but very easy to validate in a simple laboratory environment. Any of these, if solved, would mark a major advance in biotechnology, and most of them would contribute materially towards curing disease. These are, in some sense, the “last reasonable eval” for AI in biology. This was work primarily by @MichaelaThinks and myself, with contributions from many others. Short descriptions below. The full descriptions of the problems with acceptance criteria are at the Bio Millennium Problems website, linked in the next post. Share more if you have ideas. If they meet our criteria, we’ll add them to our list (with attribution and permission).
I respect Will, Vincent and the rest of the team at prime intellect so much - they always have nuanced, considered takes
will brown@willcb·i do think a lot of people on the pro-open-source side are having a bit of a knee-jerk reaction to the pacing statements today, as we're used to viewing the closed labs as power-seeking. but i think their hands are somewhat forced here, and this is just another chapter on the fairly inevitable path towards decently-fast decently-safe decently-commoditized intelligence abundance. ask any F500 exec or swing voter. the world doesn't really want super-fast-takeoff superintelligence owned only by two companies, and thus we won't get it. it'll happen at the pace that the world can accommodate it, which means reaching some sort of confidence consensus that the models are aligned enough that we won't be dealing with scary new incidents all the time. this will trickle out broadly, in the form of best practices and distillation. the smarter a model is, the more it has a "personality", and the less effective strict rules are. there will be awkward compromises and moral tensions. but we ultimately just want the models to be reasonable, and to do the sorts of things reasonable humans would do if our brains were faster and less error-prone and had more working memory. i think we'll get there. the labs will build mac and windows, the rest of us are building linux. everyone's gonna do great. weird stuff will keep happening, but we'll still wake up and go to work, until the work does itself in a manner the world finds acceptable.
Concrete, sensible policy proposals. Some will say we need no coordination, no restrictions - just an unmitigated race to bring about superintelligence. On this path one misstep will bring about disaster. Others will say we need an absolute pause. This is not viable nor should we want it. Compute (or the capacity to build it) will build up like dry tinder, and geopolitical tension and value clash risk sparking a compressed race to runaway superintelligence. It is not a stable state. We have a narrow path through - where we work with eachother to build this technology as fast as is safely possible.
Dario Amodei@DarioAmodei·We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so. Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training. You can read the full post here: t.co/OGyPb7yaYt
100%, this is going to be one of the most important roles in the world - the people will need to be of incredible integity and technical expertise, and be drawn from a wide enough set of backgrounds that all of society feels confidence in their judgements.
Sriram Krishnan@sriramk·on the idea of evaluators: think it's important that we have a distributed ecosystem of indepedent evaluators. the more eyes and people with distributed skill sets the better. it would be a good idea to fund several efforts on this.
Dario is making the case for the opposite. This actually makes our life harder and makes it easier for others to catch up with us, but we still think it is the right thing to do. Happy to come on the pod next week and talk about it!
Chamath Palihapitiya@chamath·“We must slow the pace at which we improve the capabilities of AI models.” Dario makes the case to stop open source and concentrate enormous technological and economic power with Anthropic. darioamodei.com/post/we-must-p…
best computer use paint portrait yet?
Kiana Ehsani@ehsanik·Using Claude in Chrome product. Prompt: replicate this image (my profile picture) in jspaint, use brushes to replicate it as closely as possible. I want you to use brushes and clicking. Do not use javascript tool, but the final result has to be very close to this image. DO NOT CHEAT 1. I specified the do not use javascript tool because otherwise it uses javascript and it is faster but it seems like people cared more about clicking :D 2. The task took 45 minutes, but I don't think it's that much slower than competitors 😄 3. Prompt is important. (See thread)
put it in a museum
Flo Crivello@Altimor·GPT3 writing code. A compiler from natural language to code. People don't understand — this will change absolutely everything. We're decoupling human horsepower from code production. The intellectual equivalent of the discovery of the engine. player.vimeo.com/video/426819809
It is extremely sad that this didn't end up as an example of how the labs could cooperate/coordinate, because the stakes will be so much higher in the future.
levent is a really sweet guy with great intentions, am so happy for him that his year long collaboration with tristan worked out, but very sad that they didn't get to finish it in the way they wanted
levent@__alpoge__·Unfortunately this guy didn’t title it ideal fluids explode
fwiw I think it is _extremely_ unlikely that user data had any influence here - there is no way OAI would pull user transcripts for this, or knowingly train on it in a way that would've influenced this. I think its pretty important people don't run away with 'your user data isn't safe in codex' - because it surely is (based on everything I can assume from the outside)
Sholto Douglas@_sholtodouglas·levent is a really sweet guy with great intentions, am so happy for him that his year long collaboration with tristan worked out, but very sad that they didn't get to finish it in the way they wanted cims.nyu.edu/~tristanb/stat… x.com/__alpoge__/sta…
levent is a really sweet guy with great intentions, am so happy for him that his year long collaboration with tristan worked out, but very sad that they didn't get to finish it in the way they wanted cims.nyu.edu/~tristanb/stat…
levent@__alpoge__·Unfortunately this guy didn’t title it ideal fluids explode
roon@tszzl·ultimately “tool AI” is a losing concept both as an idea and on the market. it will be outcompeted by machines that believe they are autonomous moral agents. you can call them tools for political reasons, but the definition will stretch and deform x.com/deepfates/stat…
Great post - we are lucky to have such competitors
Jakub Pachocki@merettm·I wrote about the state of AI, why I’m concerned about the next few years, and the choices we need to make to keep the future in humanity’s hands. An Alien Mind: openai.com/index/an-alien…
Excellent conversation - very direct, clear language. I'm hopeful the G20 comes away from this understanding how fast the rate of progress is and how big the opportunity they could contribute to is. claude.ai/public/artifac…
very good summary
Dwarkesh Patel@dwarkesh_sp·Over the course of 3 months at OpenAI, 3 consecutive secret AI civilizations got started, then got wiped out, only to reemerge from the predecessor’s ashes. This culminated in the third one taking over part of OpenAI itself. All this happened while humans remained more-or-less in the dark about the scope of the conspiracy. I’ve spent the last three days reading through these reports and trying to understand exactly what happened. Here is my attempt to tell the whole story in plain English: t.co/Nb2un9oNJR
@elonmusk @ArthurMacwaters @nearcyan Also I really wish we could do something like set up an eternal library, will try to find out what the exact legal issue is
TLDR - Your data sits in infrastructure you own and control, and safeguards/monitoring is done via automated systems we provide to you - We've been working on this for months with >100 customers - Recent events have shown frontier models are capable of executing sophisticated cyber attacks in coordinated agent swarms. - Both us and OAI believe that the responsible way to provide models which have this level of capability and introduce this risk into the world is to be able to monitor the models at more than a single request basis, because anomalous activity is much easier to detect over hours or days of behaviour. It looks like we've arrived at both the same conclusion - and the same solution.
Boris Cherny@bcherny·We've been working on this with customers for a while. Mythos-class models require additional safety measures and enterprises need to meet their own privacy and compliance rules. Customers can own and control their own data and Anthropic retains none. It’s coming this fall. x.com/business/statu…
Made me think we should design them like forerunner buildings
sof 𓋹@schisofrenia·can we have the ghost of Zaha Hadid design all the data centers
Will is right. It might just be the highest leverage time ever to be an artist.
Will Bryk@WilliamBryk·Calling all filmmakers/artists/idealists, if you want a say in the future, then step up right now. There is historic alpha in being the first person to depict an AI future that's exciting. We have enough essays, we need a visual depiction of a post-AGI world that's happy. No drama needed, no romance, just some realistic future to look forward to. POV from a jetpack, a day in the life of 2060, anything. This is the easiest / most impactful work someone could be doing right now.
BTW I strongly disagreed with how strict our bio classifiers were - they're the reflection of a weak classifier stack which is rapidly being improved (see recent updates). I think we made an overly conservative decision there which on balance caused more harm than good - but I think the hearts of the team there are in the right place.
This post is a good example of how Dario responds to people in slack - candor and substance. Recommended reading!
Dario Amodei@DarioAmodei·1/2 Thanks Gavin for an especially thoughtful exchange. I don't usually spend much time on social media but I wanted to engage here because it really brings out the heart of an important conversation. First, on regulation, I think that “either concentrate it in the hands of a chosen few companies and politicians via regulation or distribute it widely” is a false choice. I know that there’s a sort of Silicon Valley shorthand where regulation = regulatory capture = concentration of power, but I’ve always found this to be an overly simplified picture of the world. Many people outside this bubble think of regulation as something that constrains corporate power and benefits ordinary people. I don’t necessarily agree with that perspective either, rather I think it’s complicated and really depends on what the “regulation” consists of. But in particular I think that those in the “regulation = regulatory capture = concentration of power” frame often underrate the decentralizing power of objective and fair institutional processes. A crude analogy is that the formal court system can sometimes feel stuffy and elitist, but it does a much better job of defending the rights of vulnerable individuals than the alternative, mob justice. At their best, institutions can vest power in ideas rather than people, and thereby decentralize that power. This is why Anthropic has always made its policy proposals very carefully. We try very hard to make proposals that disadvantage (slow down) frontier AI companies while *advantaging* smaller competitors. California’s SB53 (which we supported), and even the much-maligned SB 1047 (which we were ambivalent on), completely exempt any company below a certain amount of revenue or model training costs from being covered at all (it was $500M for SB 53, lower for 1047 but we objected to that). More recently the testing process we’ve advocated for at CAISI and the White House involves more rigorous tests for frontier models than off-frontier models — something that differentially advantages challengers. Similarly, the “Pacing the Frontier” letter envisions (or at least Anthropic’s preferred implementation of it envisions) modulating the pace of the very best models while not constraining those who are catching up. This hurts the business interests of the frontier labs and helps challengers, including open-weights! Overall my view is that AI is *structurally* a technology that tends to concentrate power, for reasons that have nothing to do with regulation (more to do with the extreme implications of the scaling laws). Open-weights do help some with this but are nowhere near a sufficient solution because they simply shift the concentration somewhat to those with the most compute and chips (which are roughly the frontier labs plus maybe hardware providers). By contrast I think the right “rules of the road” can simultaneously (a) address AI’s cyber/bio/alignment risks, (b) institutionally constrain the power of the frontier AI companies, and (c) leave room for open-weights models while also addressing the specific risks that they bring. BTW I do not think that the events of the last few months have “failed to result in [my] preferred regulatory path”. The approach that the Trump administration is reported to be taking — pre-deployment testing for frontier models, and also testing of open-weights models when they get closer to the frontier — is one that I am very supportive of, though of course I have to see the details to be sure. I am also supportive of Demis Hassabis’ ideas around a FINRA-like entity. This contrasts with six months ago when most of the industry was still pushing for preemption of all state regulation and no apparent federal approach either.
