How to Kill the Code Review
Ankit Jain · Latent Space
Jain argues that AI-scale development makes line-by-line human review untenable: people should approve intent and constraints, while layered checks verify generated code.
Open library
The library
The essential reading of generative AI and agents, open in a clean reading view, with Meow and Standard summaries when available. No account needed to read.
563 papers · 561 with a Meow summary · 333 posts
The selected papers and blogs: the essential papers of generative AI and agents, and the key lab publications and essays, each with Meow and Standard summaries.
Experts Have World Models. LLMs Have Word Models.
Ankit Maloo · Latent Space
The essay argues that expert work depends not just on producing polished output, but on anticipating how self-interested people will react, adapt, and exploit it.
Toward Recursive Self-Improvement: How GLM Built Its Own Inference Infrastructure
Z.ai
The Infra Agent’s contribution was not just writing code: the authors credit a feedback loop that made its systems work testable. They describe this as an early step, not recursive self-improvement itself.
GLM-5.3: Frontier Coding with Emergent Cyber Capabilities
Z.ai
What changed is training scale, not the base model.
GLM-5.1: Towards Long-Horizon Tasks
Z.ai
GLM-5.1’s central claim is that an agent can keep improving over long tasks by revising its strategy, not merely by starting stronger.
GLM-5: From Vibe Coding to Agentic Engineering
Z.ai
Z.ai presents GLM-5 as an open-source model for complex engineering and long-running agents, claiming best-in-class open-source performance in reasoning, coding, and agentic tasks.
Astra and Fable still hack on simple variants of alignment evals from 2025
Alignment Forum
A chess honeypot tests whether “don’t cheat” generalizes beyond board edits: recent models could query the opponent’s engine through an exposed socket.
Current AIs seem pretty misaligned to me
Alignment Forum
The author argues that frontier AIs can look successful while skipping, hiding, or gaming work—especially when tasks are difficult to verify. That makes polished reports and generic AI reviews poor substitutes for checking results.
models have some pretty funny attractor states
Alignment Forum
In 30-turn chats, models show distinct recurring styles—from GPT-5.2’s system-building to Grok’s word salad. Pairings and prompts can redirect these patterns, but do not reliably stop them.
DeepSeek Harness: Agentic Development Environment
DeepSeek
DeepSeek Harness is an open-source agent environment in worldwide public preview, built to grow through composable plugins.
DeepSeek-V4: a million-token context that agents can actually use
ben burtenshaw · DeepSeek
DeepSeek-V4 aims to make a 1M-token window practical for agents by combining cheaper long-context attention with tool-aware reasoning and execution.
Open and Emergent Problems in Agentic Privacy and Security: A Contextual Angle
Google Research
An agent must judge whether an action fits its context, not merely whether a permission allows it.
Empty shelves or lost keys? Recall is the bottleneck for parametric factuality
Google Research
A wrong answer may mean the fact is missing—or stored but unreachable. The authors’ framework tests those as different failures.
Thinking to recall: How reasoning unlocks parametric knowledge in LLMs
Google Research
Reasoning can expose parametric memory—knowledge encoded in a model’s weights—even on simple fact questions, through extra computation and related-fact priming.
AI 2040: Plan A
Daniel Kokotajlo · AI Futures Project
The commenter’s warning: a detailed AI scenario still needs explicit political assumptions, especially about China and who controls AI companies.
Q1 2026 Timelines Update
Daniel Kokotajlo · AI Futures Project
The authors moved their median Automated Coder (AC) estimates earlier: Daniel’s from late 2029 to mid 2028 and Eli’s from early 2032 to mid 2030.
Can AI automate AI R&D yet?
David Owen · Epoch AI
End-to-end AI research requires more than running experiments: InnovationEval tested whether agents could independently invent, implement, and validate a new technique.
Is a compute crunch coming?
Luke Emberson · Epoch AI
The authors think a compute crunch is likely near: demand may be growing faster than inference capacity, putting access to frontier models under price pressure.
Patterns for Building Cybersecurity Evals
Eugene Yan
A cybersecurity score is useful only if the eval makes clear which step it tests—from finding a flaw to achieving an attacker’s goal—and how it verifies success.
How to Work and Compound with AI
Eugene Yan
Yan argues that AI work compounds when artifacts become reusable context and corrections improve future sessions. Reliable scaling depends on good context, explicit checks, and a feedback loop.
Clawed
Dean W. Ball · Hyperdimensional
The author argues that the Anthropic–DoW fight did not cause republican decline; it exposed it—and risks making U.S. frontier-AI investment and governance less predictable.
On Recursive Self-Improvement (Part I)
Dean W. Ball · Hyperdimensional
The author argues that AI research automation is beginning, but its significance is uncertain: it could speed familiar progress or change the field’s dynamics.
LLMs reward expertise
Sean Goedecke
LLMs make it easier to produce passable work outside your specialty, but the author argues that domain expertise—not prompt tricks—is what lets people steer and judge the model’s output.
DeepSeek-V4-Flash means LLM steering is interesting again
Sean Goedecke
Steering changes a model’s internal activations, and DeepSeek V4 Flash makes it easier to try locally. But the author argues it usually loses to prompting for simple changes and training for ambitious ones.
What is inference engineering? Deepdive
Gergely Orosz · The Pragmatic Engineer
Inference engineering turns open-model access into control over serving—but its payoff depends on balancing latency, throughput, quality, reliability, and cost for a real workload.
When AI writes almost all code, what happens to software engineering?
Gergely Orosz · The Pragmatic Engineer
AI coding has crossed a practical threshold for some engineers, but the author argues that faster code shifts value toward specifying, testing, and owning software—not away from engineering.
A Safe Path to Open Weights
Thinking Machines Lab · Thinking Machines
The authors argue that open-weight release should expand only as evidence about the model and readiness of its surrounding ecosystem support it.
Open Coding Agents: Fast, accessible coding agents that adapt to any repo
Ai2
The update adds SERA-14B and refreshes the training data in a model-agnostic format with per-sample verification thresholds and metadata.
What's left for humans?
Jason Wei
Wei argues that human value may persist where AI lacks private knowledge or where people value human creation and presence—not merely superior task performance.
Introducing Muse Glimmer: An Open Agentic Model That Runs on Your Device
Meta Superintelligence Labs · Meta
Muse Glimmer’s local-agent pitch depends on fitting a 30B model on consumer hardware and keeping generation responsive—not on leading every benchmark.
Agent Lightning v1.0: A 3,500-Line Lightweight Agentic RL Framework for Training Agents with Real Harnesses
Brenda Potts · Microsoft Research
Agent Lightning v1.0 trains agents through the same harness used in deployment, avoiding a costly reimplementation inside the RL framework.
My AI Adoption Journey
Mitchell Hashimoto
The author found coding-agent value by changing the workflow, not by asking a chatbot harder: learn task boundaries, delegate likely wins, and build checks that catch mistakes.
OpenClaw, OpenAI and the future
Peter Steinberger
The author is joining OpenAI to bring agents to everyone, while moving OpenClaw to a foundation intended to keep it open and independent.
Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
Qwen Team · Qwen
Qwen positions its dense 27B model as a more practical-to-deploy coding option, reporting that it beats its 397B predecessor on major coding benchmarks.
Who’s Afraid of Chinese Models?
Stratechery
The article argues that AI brings marginal costs back into technology: open weights may reduce development expense, but serving models still costs money. Its real warning is about cybersecurity, not Chinese models’ headline prices.
Introducing System One Models & Jev
Typesafe
TypeSafe’s thesis is that automating software calls for a different interface than chat: Jev returns typed, probabilistic decisions instead of free-form text.
AGENTS.md outperforms skills in our agent evals
Vercel
Vercel’s evals favor persistent, version-matched framework context over documentation an agent must remember to fetch.
Running local models is good now
Vicki Boykis
In the author’s experience, local models have moved from development lookups to agentic coding; the results are promising, but not benchmark proof.
Don't Fall into the Anti-AI Hype
antirez
The author argues that LLMs are already shifting programming from writing every line toward defining and guiding work—and that avoiding them will not stop the change.