Mostafa Erfani

Today, we are releasing our next milestone in building general superintelligence at Deep Cogito. We are releasing 4 hybrid reasoning models (70B, 109B MoE, 405B, 671B MoE) as a proof of concept for a novel AI paradigm - iterative self-improvement (AI systems improving themselves). Highlights: 📌  We are releasing 4 hybrid reasoning models of sizes 70B, 109B MoE, 405B, 671B MoE under open license. 📌  The largest 671B MoE model is amongst the strongest open models in the world. It matches/exceeds the performance of the latest DeepSeek v3 and DeepSeek R1 models both, and approaches closed frontier models like o3 and Claude 4 Opus. 📌  We extend our work on building superintelligence using Iterated Distillation and Amplification (IDA) by scaling the model’s intelligence prior. This is done by the model internalizing the reasoning process using iterative policy improvement, rather than simply searching longer at inference time. This is in line with the approach used for AlphaGo / AlphaZero, applied to natural language. 📌 This is a novel scaling paradigm where the models develop more “intuition”, and serves as a strong proof of concept for self-improvement. Since the Cogito models develop a better intuition of the trajectory to take while searching at inference time, they have 60% shorter reasoning chains than Deepseek R1. We are excited to get our research and models into the hands of the AI community. You can download the models on Hugging Face, or use them as an API on Together AI or Runpod or Baseten, or run them locally using Unsloth AI. We expect significant improvements by self improvement over the next few months using our approach. All models we create will be open sourced. More details in the blog below.

349

44 Comments

Prompts now sit at the heart of GenAI systems, moving through a flow that demands real engineering control. The process begins with creating and structuring prompts, adding parameters, configurations, and metadata. It continues with version control, where changes are tracked and teams collaborate. Automated evaluations, metrics, and quality gates then validate performance and reliability. Finally, prompts are integrated into production pipelines, connected to CI/CD, experiments, and monitoring. This flow is not static. Each stage feeds back into the next, enabling constant refinement and higher confidence in the system. Prompt management at this level requires engineering rigor, not one-off creativity. It is the foundation for scaling GenAI safely and effectively. What steps have you taken to make prompt management part of your engineering practice?

93

4 Comments

Now live on Vapi: GPT-5, GPT-5-Mini, GPT-5-Nano, and GPT-OSS Major releases from OpenAI this week - all available on Vapi within 1 hour of launch. → GPT-5, GPT-5-Mini, GPT-5-Nano accessible via Vapi Dashboard and API → GPT-OSS released under Apache 2.0 Read more: https://vapi.sh/ZHhWdq2

226

39 Comments

Most investors are focused on SpaceX’s momentum --> 6M+ Starlink subs, strong cash flow, and unmatched launch cadence. But Amazon is positioning itself for a familiar playbook. And it looks a lot like AWS, Prime, and Echo… – Cheaper terminals – Massive bundling potential – Prime integration – A global regulatory strategy – Deep pockets and patience Starlink has the lead today, but Kuiper is coming. 🔗 Read the full deep dive here: https://lnkd.in/eht2aAeB ...and see why investors might be seriously underestimating Kuiper.

85

4 Comments

Argon AI + OpenAI’s GPT5 unlocks the a new frontier in Life Science workflow After extensive research and testing, we found that Argon’s multi-model AI system with the integration of GPT-5 is the closest we’ve come to capturing the taste and ability of Life Science experts. Our team tested Argon’s newest Agents across Epi, Drug & Treatment Landscaping, Patient Journey, Clinical Research, and Commercial workflows. We found that Argon with the incorporation of GPT5 is a leap forward in accuracy, comprehensiveness and taste. Importantly, we found that Argon now nears full-automation of Competitive Intelligence (CI). From alert capture, to strategic deep-dives and recommendations. And we continue to explore the frontier of workflow automation along with our partners. We break down our methodology and testing, and what this means for the future of AI in Life Sciences in the comments. 

45

5 Comments

Last week we hosted a fireside chat with Thariq Shihipar, DevRel at Anthropic, on his journey from building a gaming startup to shaping Claude Code. A few insights from the evening: "The best engineers aren't getting replaced by AI, they're unblocked. Bugs that used to take half a day now vanish, and they can run three projects at once." "Good docs for humans = good docs for AI. If the content isn't there or it's outdated, no amount of formatting optimizations will save your GEO." "To build community, you just have to get over being cringe. Talk to users, even if it's just two people in a Discord. That's how it starts." Conversations like these remind me how quickly developer experience is evolving, but so much of it still comes down to fundamentals.

121

10 Comments