Topic
Coding Agents
Practical shifts in agentic coding, review, and delivery. Curated tldw.news briefings about coding agents, with practical engineering takeaways from long-form AI and developer-tool videos.
49
breakdowns
Page 1 of 5
AI EngineerWhy most AI agent benchmarks are lying about 'long-horizon' capability
Most AI agent benchmarks claim 'long-horizon' capability but measure tasks with minimal state dependency. Theta Software explains why this distorts adoption…
AI EngineerPost-training shifts from synthetic environments to messy production learning
Post-training is moving from synthetic environments to real production harnesses. The tradeoff: controlled RL vs. messy but realistic learning. Reward…
AI EngineerAI That Optimizes Its Own Kernels: Real Progress or Hype?
Recursive AI claims their system outpaced human experts on CUDA kernel optimization. But the line between automated research and recursive self-improvement…
AI EngineerGroup agents need new security, memory, and privacy playbooks
Single-user agents are easy. Group agents aren't. New security, memory, and privacy challenges emerge when an agent serves multiple users.
Hugging FaceGRPO for LLMs: Reward Design Matters More Than Algorithm Choice
GRPO makes RL for LLMs accessible, but reward hacking is a real risk. The key is group variation and careful monitoring—not just watching reward curves go up.
AI EngineerNetflix’s AI agent playbook: Stop fixing performance manually, build a pattern catalog
Netflix engineers built AI agents that turn profiling data into performance fixes in minutes. The key is a reusable pattern catalog that shifts optimization…
Y CombinatorBuilding on LLMs: delete your system prompt, let the model run
Claude Code's creator reveals why you should delete your system prompts and give models harder tasks. The real skill is elicitation, not prompt engineering.
David OndrejCode is dead. Long live workflow design for AI agents.
AI agents are making code cheap. The real work is now designing workflows and information systems. Forston Ball explains why local dev is dying and how to…
Theo - t3․ggOpus 5: The first practical default model for AI-assisted coding
Opus 5 offers a compelling middle ground between capable and cheap coding. Real savings are 20-25%, not 50%. Teams should test it as a daily driver before…
Y CombinatorNo Code In 2 Quarters: Datadog's AI Pivot
Datadog says AI will end code writing in 2 quarters, after devs rebuilt 6-month systems in days, forcing reorg and tough questions.
Get TL;DW
Too Long; Didn't Watch.
A concise breakdowns of the AI and devtools videos that actually matter for engineering leaders.
Free. Weekly. No hype.