Francesco Mosconi

As someone who usually disagrees with @DarioAmodei, I gotta give him *major* props for this one. “If you believe you have the right to force me—use your guns openly. I will not help you to disguise the nature of your action." This is what Hank Rearden, the steel magnate from Ayn Rand’s Atlas Shrugged, tells the government after disobeying its command to run his business in a manner contrary to his conscience. Amodei channeled Rearden’s spirit in a letter refusing to remove safeguards from Claude, as requested by the War Department.

Today feels like a turning point in motivation for why I do what I do. Deciding to drop everything else and focus my efforts on AI safety felt right then and feels even more right today as the stakes get real. Proud to work here!

I've updated towards a higher chance of much faster than trend AI progress in 2026. I also generally updated towards (somewhat) faster doubling times in the next 2 years. I was expecting ~170 day doubling times (5.7 months) and now I expect ~150 day doubling times (5 months).

@chrisbarber @mikeyk @calvinfo In my case the choice primarily came from a desire to work on AI safety.

I made a list of ex-founders now working at big AI labs as non-founder employees. You’re probably aware of @mikeyk (IG co-founder) at Anthropic. You might’ve also seen @calvinfo, co-founder of Segment ($3.2B acquisition), joined OpenAI and recently moved on. The List: @vedantmisra, founder of Kemvi, acq by HubSpot, now Gemini core team at GDM @framosconis, GP at Pioneer Fund, now at Anthropic @BohanZhangOT, cofounder of OtterTune, now at OpenAI @casparro who has started some interesting self funded arbitrage businesses, now at OpenAI @silber and @michael_dworsky, cofounders at Cove, now at Anthropic @tibbetts, cofounder at Streambase (acq TIBCO), now at Anthropic Kanav Kalucha, cofounder at Amarsha (YC W24), now at OpenAI Anuj Sharma, cofounder at Decrypt Bio, now at GDM Niki Parmar (Attention paper), cofounder at Adept & Essential, now at Anthropic @JackLZhu, did his own prop quant trading, now at Anthropic @GuruDetective, Bits N Bytes (nonprofit taught 20k+ students), now at Anthropic @alexkarpenko, cofounder of Luma Camera (acq IG) and Rylo (acq VSCO), now at OpenAI @_enochc, cofounder of Prelay, now at OpenAI @aslvrstn, cofounder of a quant trading firm, now at Anthropic @markjen, cofounder of Common Networks, now at Thinky @z, another cofounder of Common Networks, now at OpenAI @stevendcoffey, cofounder of Trava, now at OpenAI @MohitReddy13, cofounder of Fennel (acq Databricks), now xAI @tall, creator of stableboost a popular early SD tool, now at OpenAI @neelmango, cofounder of GameSnacks (acq Google), now at OpenAI @chakmingli, cofounder Storm8, now at OpenAI @chengyinliu, cofounder Slab, now at Anthropic @ajambrosino, cofounder Catch, now at OpenAI Abe Land, cofounder Ardius (acq Gusto), now at OpenAI @giansegato, cofounder Uniwhere, now at Anthropic @realtsmc, cofounder Rootbeer Computer, now at OpenAI Nikhil Bhargava, cofounder Metamanagement, now at Anthropic @sagunamatata, did their own consumer product studio, now at OpenAI Scott Zhang, cofounder SamaCare, now at OpenAI Raphael Hoffmann, cofounder Lattice, now at GDM Samuel Wookey, cofounder Masterful, now at OpenAI @theteriyu, cofounder Vibely (acq Kajabi), now at OpenAI @ralucaadapopa, cofounder Opaque Systems, now at GDM @eschoppik, cofounder Rithm School, now at Anthropic Peter Lai, cofounder Emburse, now at Anthropic Abhijeet Dwivedi, cofounder ZeroDown and Central, now at OpenAI @newhouseb, cofounder AidKit and bubbli, now at OpenAI @mwil20, cofounder Prufrock Ventures and Swithboard, now at OpenAI Frank Sifei Luan, cofounder Wayo, now at xAI Dan Roberts, cofounder Diffeo, now at OpenAI Then there are the acquisitions too: @_mohansolo & Douglas Chen, founders of Windsurf (acq GDM/Cognition), at GDM @RazRazcle, Peter Hayes, and @jordnb, founders of Humanloop (acq Anthropic), at Anthropic @alexandr_wang from Scale at MSL @natfriedman (Ximian and Xamarin) at MSL @danielgross (SSI, Pioneer and Cue/Greplin) at MSL @turtlesoupy, Taylor Gordon and @wjosephflynn, cofounders of Global Illumination (acq OpenAI) at OpenAI @iamveeve and @dhruba_rocks and @tudor from Rockset (acq OpenAI) at OpenAI @potatoarecool @embirico from Remotion (acq OpenAI), at OpenAI Lmk the main ones I missed. Side note, joining a lab as an ex-founder seems like a good move. 1. rare knowledge you can utilize for your future startups 2. preview the future of ai 3. observe how to grow models 4. crazy talent density for future cofounders 5. labs need your skills, they have distribution, and lots of tech innovation, that needs to be productized and delivered to users

Check out the system card addendum here: assets.anthropic.com/m/4c024b86c698…

We conducted voluntary automated testing to track capability progression and validate our safety assumptions.

Today we released Claude Opus 4.1! Our most powerful model yet.

So many things to love about Claude 4! My favorite is that the model is so strong that we had to turn on additional safety mitigations according to Anthropic's responsible scaling policy

Proud to announce that for much of this year Anthropic has been working with the National Nuclear Security Administration (NNSA, part of DOE) to test out whether LLMs like Claude know about dangerous things relating to nuclear weapons. axios.com/2024/11/14/ant…

Proud to keep working on making the most advanced AI safe!

Late to share, but absolutely proud of this work towards making AI safer and more interpretable!

@theficouple Where do you still find rental properties that can not only cover your mortgage but also give you income, given current interest rates?

1. Housing: The average person spends almost 27% of their income on housing. We didn’t want average so we chose to house hack. This meant buying rentals & using income from them to cover our rent. This gave us $0/mo housing.

Today we spend just under $6,000/mo as a family of 3. But from 2020-mid 2023 our bills were ~$2,300. During this time, our income tripled which helped us save/invest more each month than we spent on bills. Here’s 3 tricks we used to cut our expenses:

@danfaggella Haha you have an article ready for everything. Nice!

@danfaggella Not sure about this. People make me discover new things. Ai will too but if it only does what I tell it will get very boring very soon. The echo chamber works because it’s imperfect. Make it perfect and it will break

"I'm going to use #AI to generate some really cool movies/music that people will love!" ^ That creator phase will be VERY short-lived. Soon, ppl will use prompts/preferences to conjure their OWN movies/music, hyper-personalized to their own needs. They won't want your stuff.

@danfaggella True. Would your definition of perfect require biofeedback? In the husk it may work

@framosconis You see perfect to be repetitive and optimized in a single way I see perfect to be optimally calibrated to the varying interests and attention of each individual user. We are defining perfect in a different way. See you in the husk sir: danfaggella.com/husk

Yes. In hindsight it was obvious.

If you’re surprised by the speed of AI development you’re not paying attention.

You want AI copilots to do what you don’t want to do to free you up to do what you are uniquely good at… … until they are good at everything you could possibly do 🤔

Generative AI makes creating these kind of images extremely easy. How will we protect ourselves from embedded subliminal messages everywhere?

The black clothes form the subliminal message: "OBEY". You can find the original reddit post here: reddit.com/r/StableDiffus…

This image is all around twitter, and it took me a while to understand why. Step away from the screen, and step back until you notice it, after that you can't unsee it.