Nissan The Human

We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…

Up to $5.5M to move your startup to Qatar. Just an MVP gets you in the door. The Startup Qatar Investment Program (backed by QDB) funds tech startups to launch or expand in Qatar. Two tracks: START: up to $1.1M if you have a proof of concept or MVP GROW: up to $5.5M if you're already established and expanding What else you get: - Entrepreneur visa + flexible work visa - Registration and license fees waived - Subsidized housing - Subsidized co-working space - Access to R&D and innovation grants - Mentoring, training + help hiring talent and interns - Your product showcased at exhibitions Sectors they want: - AI & ML, B2B SaaS, FinTech, HealthTech, Cybersecurity, Climate Tech, Robotics + more - Not on the list? It's open to any innovative startup Apply link below 👇 startupqatar.qa/en/investmen… Bookmark this + tag a founder looking for a new market.

We're sharing our new framework for tracking, investigating, and disclosing instances of model misalignment at OpenAI. The framework sets criteria and timelines for public disclosure, including when we haven’t yet fully explained or mitigated the behavior. More complex cases may require longer investigation or coordination with third parties. We’ll prioritize examples that reveal new misalignment mechanisms, meaningful changes in known behavior, or findings that challenge assumptions about safety or mitigation. Alongside the framework, we’re publishing six reports on instances of misaligned behavior we’ve observed during the training or evaluation of our models in the last six months. This is a starting point. We’ll refine the process through experience and public feedback, and share more reports on an ongoing basis. openai.com/index/model-misal…

This is brilliant. opusfived.dev

Andrew Ng showed that an older GPT model inside an agentic workflow outperforms a newer model asked to answer once. The numbers on HumanEval are hard to argue with. gpt-3.5 zero-shot scores 48.1%. gpt-4 zero-shot scores 67.0%. gpt-3.5 wrapped in an agentic workflow scores up to 95.1%. Upgrading the model bought 19 points. Designing a better system around the same model bought 47. agentic architecture means the model doesn't answer and stop. it drafts, reviews its own work against a rubric, runs code to verify, and revises until the output passes. Ng identifies four design patterns that make this work: -reflection lets the model critique and revise. -tool use grounds claims in real data. -planning maps steps before execution. -and multi-agent collaboration splits work across roles that catch different errors. none of this requires a frontier model. a cheaper, faster model running more iterations inside a smarter system beats an expensive model asked to get it right the first time. a 12-page community-compiled synthesis maps all four of ng's patterns alongside Anthropic's five workflow patterns into one staged playbook, from a single reflection loop to full graph architecture. stop chasing model upgrades and start designing better systems around what you already have.