blog/
107 pages · Updated July 18, 2026
Pages
- From bare metal to a 70B model: infrastructure set-up and scripts - Imbue
- A PyTorch implementation of SIMONe - Imbue
- Imbue Blog: Writings about honest (and punk!) software - Imbue
- Understanding self-supervised and contrastive learning with "Bootstrap Your Own Latent" (BYOL) - Imbue
- VICReg: Tutorial and lightweight PyTorch implementation - Imbue
- A case study in testing with 100+ of Claude agents in parallel - Imbue
- Scaling Laws For Every Hyperparameter Via Cost-Aware HPO - Imbue
- Usefully run hundreds of Claudes in parallel with mngr - Imbue
- Jack Parker-Holder, DeepMind: On open-endedness, evolving agents and environments, online adaptation, and offline learning - Imbue
- Fine-grained HTTP permissions for AI agents - Imbue
- The state of LLM ethical decision-making - Imbue
- Noam Brown, FAIR: On achieving human-level performance in poker and Diplomacy, and the power of spending compute at inference time - Imbue
- Ari Morcos, DatologyAI: On leveraging data to democratize model training - Imbue
- Empowering humans in the age of AI - Imbue
- Sugandha Sharma, MIT: On biologically inspired neural architectures, how memories can be implemented, and control theory - Imbue
- Imbue raises $200M to build AI systems that can reason and code - Imbue
- Sergey Levine, UC Berkeley: On the bottlenecks to generalization, why simulation is doomed to succeed, and picking good research problems - Imbue
- Julian Chibane, MPI-INF: On 3D reconstruction using implicit functions - Imbue
- Martin Arjovsky, INRIA: On benchmarks for robustness and geometric information theory - Imbue
- Jonathan Frankle, MIT: On the lottery ticket hypothesis and the science of deep learning - Imbue
- Introducing Generally Intelligent - Imbue
- Hattie Zhou, Mila: On supermasks, iterative learning, and fortuitous forgetting - Imbue
- Our approach to safety - Imbue
- There Will Be a Scientific Theory of Deep Learning - Imbue
- Closed vs. open AI ecosystems: Kanjun on Decoder - Imbue
- The AI paradox: Kanjun at the 2024 Progress Conference - Imbue
- Developing agents that can reason and code: Kanjun and Josh on the No Priors podcast - Imbue
- Human agency vs. “agentic” AI: Kanjun on the NonZero podcast - Imbue
- The future of AI is human-centered: Kanjun on the Pioneers of AI podcast - Imbue
- Joel Lehman, OpenAI: On evolving intelligence, open-endedness, and reinforcement learning - Imbue
- A more empowering view of general agents: Kanjun on NYSE TV - Imbue
- Agency in the age of AI: Kanjun on the Minus One podcast - Imbue
- A Healthy Ecosystem for AI Agents: Introducing the AFI - Imbue
- LLM-based Evolution as a Universal Optimizer - Imbue
- Open-sourcing CARBS: how we used our hyperparameter optimizer to scale up to a 70B-parameter language model - Imbue
- Sarah Jane Hong, Latent Space: On neural rendering & research process - Imbue
- Beating ARC-AGI-2 with Code Evolution - Imbue
- Glenn McDonald: On Dungeons & Dragons, data alchemy, and radically open software - Imbue
- Appendix for "Understanding self-supervised and contrastive learning with 'Bootstrap Your Own Latent' (BYOL)" - Imbue
- Where are the good policy ideas? Analyzing the NTIA's AI RFC - Imbue
- Bryden Fogelman: On coding agents and finding meaningful work at Imbue - Imbue
- Keystone-Eval: Our agentic benchmark for self-configuring agents - Imbue
- How Offload Works: Inside the Rust CLI that sped up our tests by 6x - Imbue
- Offload: We made agents run our test suite 6x faster - Imbue
- Every mngr plugin, explained - Imbue
- How AI code review can make correct code worse - Imbue
- Drew Linsley, Brown: On inductive biases for vision and generalization - Imbue
- Methodology for analyzing the individual comments to the NTIA’s AI RFC - Imbue
- Dylan Hadfield-Menell, UC Berkeley/MIT: On the value alignment problem in AI - Imbue
- Training a 70B model from scratch: open-source tools, evaluation datasets, and learnings - Imbue
- A more radical Imbue - Imbue
- The need for trust: Analyzing the individual comments to the NTIA’s AI RFC - Imbue
- Yujia Huang, Caltech: On neuro-inspired generative models - Imbue
- Towards guardrails, not guidelines: a policy framework for powerful AI systems - Imbue
- Blueprint: One-shot bigger coding tasks - Imbue
- Seth Lazar, Australian National University: On legitimate power, moral nuance, and the political philosophy of AI - Imbue
- On the Stepwise Nature of Self-Supervised Learning - Imbue
- Yash Sharma, MPI-IS: On generalizability, causality, and disentanglement - Imbue
- Choices and Knives - Imbue
- Chelsea Finn, Stanford: On the biggest bottlenecks in robotics and reinforcement learning - Imbue
- Rylan Schaeffer, Stanford: On investigating emergent abilities and challenging dominant research ideas - Imbue
- How to usefully run 1,000 agents in parallel - Imbue
- Tri Dao, Stanford: On FlashAttention and sparsity, quantization, and efficient inference - Imbue
- Sculptor: the missing UI for parallel coding agents - Imbue
- Celeste Kidd, UC Berkeley: On attention and curiosity, how we form beliefs, and where certainty comes from - Imbue
- Jim Fan, NVIDIA: On foundation models for embodied agents, scaling data, and why prompt engineering will become irrelevant - Imbue
- A PyTorch implementation of slot attention - Imbue
- Oleh Rybkin, UPenn: On exploration and planning with world models - Imbue
- Percy Liang, Stanford: On the paradigm shift and societal effects of foundation models - Imbue
- Bouncer: Heal your feed - Imbue
- Vincent Sitzmann, MIT: On neural scene representations for computer vision and more general AI - Imbue
- What's next in AI?: Kanjun and Reid Hoffman at Config 2023 - Imbue
- Why AI agents don’t work (yet): Kanjun on the Latent Space podcast - Imbue
- One-person enterprise: Kanjun at the World Economic Forum Annual Meeting - Imbue
- Nicklas Hansen, UCSD: On long-horizon planning and why algorithms don't drive research progress - Imbue
- 'We need legal protections from AI’s risks': Matt Boulos in The Hill - Imbue
- How we made sandboxed coding agents 10x faster to start - Imbue
- Our technical hiring process - Imbue
- Jamie Simon, UC Berkeley: On theoretical principles for how neural networks learn and generalize - Imbue
- Jacob Steinhardt, UC Berkeley: On machine learning safety, alignment and measurement - Imbue
- Andrew Lampinen, DeepMind: On symbolic behavior, mental time travel, and insights from psychology - Imbue
- Minqi Jiang, UCL: On environment and curriculum design for general RL agents - Imbue
- Giancarlo Kerg, Mila: On approaching deep learning from mathematical foundations - Imbue
- Imbue Raises Fresh Capital, Amazon's Alexa Fund Joins Investor Lineup - Imbue
- Transforming AI agents into personal collaborators: Kanjun on the NVIDIA AI Podcast - Imbue
- Yilun Du, MIT: On energy-based models, implicit functions, and modularity - Imbue
- Avalon: A Benchmark for RL Generalization Using Procedurally Generated Worlds (NeurIPS Spotlight Paper) - Imbue
- Our submission to the NTIA's AI accountability RFC - Imbue
- Reinventing AI agents: Kanjun on the Gradient Dissent podcast - Imbue
- Latchkey: Use plain old REST instead of MCP - Imbue
- Archit Sharma, Stanford: On unsupervised and autonomous reinforcement learning - Imbue
- Update: The state of LLM ethical decision-making - Imbue
- Training >70B LLMs on 10,000 H100 clusters: Josh on the Latent Space podcast - Imbue
- Jupyter Ascending: Sync notebooks to your editor - Imbue
- Ben Eysenbach, CMU: On designing simpler and more principled RL algorithms - Imbue
- Use your unused Claude tokens for good - Imbue
- Bill Thompson, UC Berkeley: On how cultural evolution shapes knowledge acquisition - Imbue
- The ethical and practical future of AI: Matt Boulos at the MIT Club of Northern California - Imbue
- Vet: Prevent coding agents from making mistakes - Imbue
- Introducing Keystone: Building self-configuring agents - Imbue
- Katja Schwarz, MPI-IS: On GANs, implicit functions, and 3D scene understanding - Imbue
- Sculptor: Catch and fix issues as you code - Imbue
- How to build robust software systems with AI: Josh at the AI Engineer World's Fair 2025 - Imbue
- Keeping humans in the loop: Josh at the Democratize Intelligence Summit - Imbue
- Generally Intelligent #0: A podcast for deep learning researchers - Imbue
- Digital Freedom Depends on Access Rights: Matt Boulos in Lawfare - Imbue
- Cinjon Resnick, NYU: On activity and scene understanding - Imbue