Head of Foundation Models @Salesforce 🤖 Founder @SpindleAI (acquired @Salesforce) 🛠ex-President Brainspace AI (acquired, Reveal) ex-CTO Volley ML (@JPMorgan)
SF
Joined April 2010
- Tweets11.7K
- Following118
- Followers1.1K
- Likes10.1K
Carson Kahn retweeted
Introducing SubQ Snap, the first long-context decision model for real enterprise work.
Snap is the only decision model or LLM where compute per token stays constant as context grows.
• 384K tokens of multimodal context, 12x larger than Jev.
• 61.7 on Decision Index v0.2.1 vs Jev’s 57.9, outperforming on 13 of 16 enterprise benchmarks.
•Processes 100K tokens and returns a decision in under one second.
•$0.05/MM input tokens. Output is free.
Available today to anyone.
The first 10,000 sign-ups get 100M free input tokens.
We’ll likely hit capacity fast and will add users as we add compute. Compute offers welcome 🙂.
Latency will be variable as we tune the platform to this new model and category of users.
Model speeds are 80ms for minimal context, 250ms for 32k, 660ms for 100k, and 2.5 seconds for 384k.
Visit: subq.ai/
Carson Kahn retweeted
Training a model to generate RL tasks not too hard, not too easy costs many solver runs per task.
PROPEL predicts difficulty via a probe on its activations instead, amortizing cost and speeding up generator optimization.
New open-ended RL research from @Vmax + @GoodfireAI.
Carson Kahn retweeted
Wrote a short blog about the "shape" of language models, and the tradeoffs they may present in the future. I genuinely think it's a valuable research direction to start thinking about now, especially w.r.t. harness design.
alexzhang13.github.io/blog/2…
Huge congrats to @ItsAlfredW, @Florian_Jue, and the entire @ListenLabs team! Grateful for this opportunity to build with such a talented crew.
AI that can understand -- and even simulate -- what customers will do can help generate better products, curate context for agents, and connect companies & governments to their customers in a whole new way:
salesforce.com/news/stories/… – at San Francisco, CA
Carson Kahn retweeted
And ES might be equivalent to gradient descent
nature.com/articles/s41467-0…
Carson Kahn retweeted
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI?
I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev
• 20-200x faster
• 40-400x cheaper (w/ output tokens free)
• Frontier composable intelligence optimized for decisions
AFAICT the shortest path to AI-based economic revolution
Self-improving agents are the next frontier: “The agents that win the next few years won’t just be ones with the cleverest foundation model,” @RyanAtallah and I explain, “they’ll be the ones that learn from their own outcomes.”
@SFResearch was among the industry’s first to demonstrate a reinforcement learning (#RL)-like model for optimizing agents, called Retroformer. At @Salesforce AI, we’re not just testing new experiments with various @Agentforce design partners ‑‑ we’re dogfooding them ourselves!
Learn more in our new 9-minute #explainer about recursive self-improvement (#RSI) for enterprise agents: salesforce.com/news/stories/… – at San Francisco, CA
Many thanks to…
• Our team from @SpindleAI, acquired by @Salesforce in 2025
• The machine learning platforms team under @Derya_Isler
• The @Agentforce team led by @MadhavTT, @JustJayesh, @NotHackedJK, Christina Abraham, Phil Mui, Kishan Chetan, @Moonhub’s Nancy Xu, and @Sanjna_Paru
• The AI Research team under @SilvioSavarese and @ItaiAsseo
• The technology organization led by @STallapr23 and MK
…for making these experiments possible: sfdc.co/self-optimizing-agen… – at San Francisco, CA
![The self-improvement loop is a flywheel inspired by reinforcement learning (RL), in which the agent powered by a frontier model continuously improves through interconnected cycles of autonomous observation, diagnosis, simulation, and (optionally) human input. This is a compound version of recursive self-improvement, or RSI, that allows agents to optimize or evolve toward improved outcomes semi-autonomously. [Agentforce, Salesforce AI]](https://pbs.twimg.com/media/HN7--DkaoAAU_qH.jpg?name=orig)
ALT The self-improvement loop is a flywheel inspired by reinforcement learning (RL), in which the agent powered by a frontier model continuously improves through interconnected cycles of autonomous observation, diagnosis, simulation, and (optionally) human input. This is a compound version of recursive self-improvement, or RSI, that allows agents to optimize or evolve toward improved outcomes semi-autonomously. [Agentforce, Salesforce AI]
Carson Kahn retweeted
first open weight thinking machine model!! 975B total, 41B active trained on 45T tokens, 1M context, multimodal in
sliding window with a 5:1 ratio and 512 size, deepseek aux-free load balancing and 2 shared experts (usually people only use 1), actually curious why the model is less sparse than kimi (~4.2% vs 3.2%). they use a short convolution after k and v, output and ffn (see plot), muon (they cite manifold muon but mention weight decay so not sure), muP, and have a very nice RL scaling curve and chain of thought!
one very cool part of the release imo is how well their small variant (276B total, 12B active) performs compared to the big one. they mention they changed the pre-training data mix and recipe, very curious about those changes and to see them scaled up to ~1T (or more?) soon đź‘€
> "It is not the most performant model available today, closed or open. We trained Inkling for solid capabilities across the board rather than state-of-the-art performance in a single area, to serve as a foundation for the models we will train in the future."
also this is really refreshing to see in a model release, huge congrats :)
Today, we are introducing Inkling.
Inkling reasons efficiently across text, image, and audio modalities. We are making the full weights available.
thinkingmachines.ai/news/int…
Available today for fine-tuning on Tinker. Play with it in the Inkling Playground. 🧵
Incredible foresight about the scaling laws of LLMs from @JeffDean et al… all the way back in 2007 🤩
“Significantly, we found that [output] quality … continues to improve with increasing language model size, at even the largest sizes considered. This finding underscores the value of being able to train and apply very large language models, and suggests that further performance gains may be had by pursuing this direction further.” (Source: aclanthology.org/anthology-f…)
Anyway, @thinkymachines co-founder @lilianweng has done a beautiful job distilling the essence of deep learning, as always -- worth a reread:
lilianweng.github.io/posts/2…
Announcing the next chapter: @SpindleAI has entered into a definitive agreement to be acquired by @Salesforce, accelerating the @Agentforce roadmap and strengthening the future of enterprise AI for millions of global users. 🤖
This is a great outcome for our team & investors. 🙏 When my cofounder Ryan and I launched Spindle in late 2021 to deliver #AI-native business analytics and scenario intelligence for the Fortune 500, no one (not even us!) fully anticipated how transformational #agents & AI-native data models like ours would be.
But with the support of…
🛠️ The smartest, most creative, and most devoted team I’ve ever had the privilege of working with
🏦 Backing from top-percentile investors like @Accel and others across Sand Hill
🧙World-class advisors like ex-@GoldmanSachs CTO Dr. Paul Walker, ex-@DeutscheBank CTO Kim Hammonds (who I miss dearly), ex-Adaptive Insights founder Rob Hull, and ex-@Sequoia partner Bob Cohn
🤖 Visionary customers, from @IBM @Apptio to @1Password and beyond
…sci-fi has become sci-fact! 🚀 Joining #Agentforce will kick Spindle AI’s journey to the next level.
I’m deeply grateful for my co-founder @RyanAtallah. Ever since we bonded in high school, he has always been not just the hardest-working leader I know, and not just the most thoughtful, but also the most loyal & trustworthy. Find a cofounder like that: their traits are echoed in the organizations they build, and the teams they invest in.
Uncountable thanks to the many dear friends, teammates, and colleagues, who have worked so hard to make this journey the privilege of a lifetime. 🙏
Most of all: a lifetime’s thanks to my wife Seda and Ryan’s wife Ashly. Without their incredible support, sacrifices, and contributions, none of this would have been possible. 🤍
My life’s mission has been machine learning that augments human cognition. With the acquisition of Spindle AI, that mission continues at Agentforce: salesforce.com/news/stories/…
Carson Kahn retweeted
Salesforce has signed a definitive agreement to acquire @SpindleAI, a leading agentic analytics platform.
The acquisition brings aboard deep expertise in multi-agent systems and high-performance data applications to accelerate Agentforce: sforce.co/4hLOpRA
Carson Kahn retweeted
Learn to ship. Shipping is a skill distinct from coding. Shipping is designing, coding, QAing, story-telling, teaching, marketing, selling, pivoting, iterating…
It used to be that coding dominated in importance because of coding ability scarcity. AI will push you to go further.
Carson Kahn retweeted
Happy to announce that spindle.ai has accepted an invitation to become a @Microsoft AI Cloud Partner!
As former @GoldmanSachs CTO and enterprise #AI leader Dr. Paul Walker notes, “Spindle has uniquely delivered agentic AI for our hardest planning and business analytics problems, and I’m thrilled to see this technology become more accessible to Microsoft’s ecosystem of customers.”
With contributors from both @MicrosoftAI and @SpindleAI having been at the forefront of #agents since their respective early work on the #GPT family of LLMs, we’re now opening the door to bring rigorous, agentic business analytics into the platforms enterprises already rely on — @Microsoft365 & @Azure.
“We’re proud that the trust of enterprises embodied in our stage-leading SOC 2 Type II attestation can now extend globally through Microsoft’s class-leading AI Cloud Partner program,” adds Spindle co-founder & AI leader @CarsonKahn, “giving customers confidence that agentic business analytics is secure, transformative, and enterprise-ready.”
👉 More big things ahead — find our teammates throughout @Workday’s #WDAYRising conference this week (lnkd.in/ghHUfDBD) for a preview of what’s to come. If you’re attending, reply here so we can connect! – at San Francisco, CA
🚀 Introducing AI agents that make “self-driving” machine learning usable by anyone — no PhD required. A lifelong dream for me, launched in months… Demo below!
Forecasting, key driver analysis, predictions, tipping-point analysis, and anomaly detection now take minutes instead of weeks. 🤖
✨ To my knowledge, we’re the first team in the world to ship autonomous machine learning (AutoML) guided by an agent — or “agentic AutoML” — usable by anyone.
Genuinely in awe of the creativity, ingenuity, and grit every @SpindleAI teammate poured into making that possible. 🙏
With transparent, personalized explanations for “how”, “why,” and “what next,” the new agent also delivers actionable recommendations for setting & hitting business targets 🎯 based on patterns in your data.
From churn prediction & customer behavior forecasting, to pricing simulations & cross-sell targeting, beta customers are already sharing compelling use cases with us…
But most of all, I’m excited to see all the creative use cases YOU come up with: spindle.ai/agents/autonomous…