tag:news.letsmake.ai,2026-11-01:feed Daily digest of news on AI as an engineering tool. Summaries are machine-generated. 2026-11-03T07:24:15Z letsmake.ai news tag:news.letsmake.ai,2026-11-02:digest 2026-11-03T07:24:15Z

19 stories for 2 November 2026.

tag:news.letsmake.ai,2026-11-02:sparks-of-weight-space-programming-llms-can-natively-read-wr 2026-11-03T07:22:54Z

Frontier LLMs like Astra can interpret and modify neural network weights without training or gradients, improving performance in games and math tasks. This suggests a new paradigm for weight-space programming.

tag:news.letsmake.ai,2026-11-02:llm-0-37 2026-11-03T07:22:59Z

LLM 0.37 includes new models, improved embedding support, and AI-assisted PR triage, with a new default model that's half the cost of the previous one.

tag:news.letsmake.ai,2026-11-02:ainews-llms-can-read-write-neural-network-weights-weight-spa 2026-11-03T07:23:06Z

A new technique allows LLMs to directly read and edit neural network weights, bypassing traditional code-based training. This could enable autonomous model improvement without human intervention, though it's limited to simple tasks for now.

tag:news.letsmake.ai,2026-11-02:veriharm-bench-measuring-harmful-success-in-llm-coding-agent 2026-11-03T07:23:11Z

VeriHarm-Bench is a new benchmark that measures harmful success in LLM coding agents, where agents complete tasks correctly but cause unintended harm. It shows that even high-performing models like GPT-5.6-Sol and Claude-Opus-5 struggle with safety.

tag:news.letsmake.ai,2026-11-02:cove-a-task-adaptive-code-based-ensemble-for-proactive-verif 2026-11-03T07:23:15Z

COVE is a new tool that improves the accuracy of LLM outputs by preventing errors before generation and validating results afterward, without requiring extra LLM calls.

tag:news.letsmake.ai,2026-11-02:cr-bench-v2-evaluating-the-real-world-utility-of-ai-code-rev 2026-11-03T07:23:20Z

CR-Bench-v2 is a new benchmark for evaluating AI code review tools, showing that even the best systems detect only about one-third of known defects.

tag:news.letsmake.ai,2026-11-02:apg-agentic-static-analysis-infrastructure-on-llvm-ir-graph 2026-11-03T07:23:25Z

APG is a new tool that enables LLM-based coding agents to perform static analysis on C/C++ code by leveraging LLVM-IR graph databases. It provides 38 atomic capabilities for analysis and repair, outperforming existing methods in benchmark tests.

tag:news.letsmake.ai,2026-11-02:built-on-broken-assumptions-boba-why-coding-agents-cannot-be 2026-11-03T07:23:30Z

A new benchmark shows coding agents struggle with chained requests, completing only 51% of tasks when requests are linked, compared to 95% when handled individually.

tag:news.letsmake.ai,2026-11-02:swe-trail-history-guided-trajectory-collection-for-coding-ag 2026-11-03T07:23:34Z

SWE-Trail is a new method for collecting coding agent trajectories that uses past data to guide future collection, improving performance without exposing history to the model during training.

tag:news.letsmake.ai,2026-11-02:veribase-automated-construction-of-theorem-libraries-for-sof 2026-11-03T07:23:38Z

VeriBase is a framework that automatically builds reusable theorem libraries for software verification by generalizing code and theorems from successful verification tasks. It improves success rates and reduces proof completion time.

tag:news.letsmake.ai,2026-11-02:crossswe-benchmarking-cross-repository-joint-development-for 2026-11-03T07:23:42Z

CrossSWE is a new benchmark testing coding agents' ability to make coordinated changes across multiple repositories. It shows that even the best models struggle with cross-repository consistency, achieving only 58.48% Pass@1.

tag:news.letsmake.ai,2026-11-02:launch-hn-momentic-yc-w24-agents-that-catch-bugs-by-using-yo 2026-11-03T07:23:46Z

Momentic is an AI QA tool that tests apps end-to-end by using the app itself to verify changes, reducing manual testing time and improving reliability.

tag:news.letsmake.ai,2026-11-02:quoting-kyle-kingsbury 2026-11-03T07:23:50Z

Kyle Kingsbury argues that the use of large language models (LLMs) undermines trust in software documentation and systems. Maintainers must find ways to ensure accuracy and demonstrate quality in an era where documentation can be easily faked.

tag:news.letsmake.ai,2026-11-02:scaling-git-at-openai 2026-11-03T07:23:53Z

OpenAI developed Pando, a Git-compatible backend system to handle the scale of development required by Codex.

tag:news.letsmake.ai,2026-11-02:qodo-builds-context-into-ai-code-review-with-codex 2026-11-03T07:23:56Z

Qodo enhances AI code review by incorporating company-specific knowledge and coding standards into Codex, enabling developers to catch critical issues before merging code.

tag:news.letsmake.ai,2026-11-02:distdebug-bench-can-llm-agents-diagnose-bugs-in-distributed 2026-11-03T07:24:00Z

A new benchmark called DistDebug-Bench evaluates LLM agents' ability to diagnose bugs in distributed systems, revealing significant gaps in their current capabilities.

tag:news.letsmake.ai,2026-11-02:can-coding-agents-sustain-continual-evolution-without-forget 2026-11-03T07:24:04Z

A study shows that coding agents can lose performance on earlier tasks when evolving to new ones, highlighting the challenge of continual self-evolution without catastrophic forgetting.

tag:news.letsmake.ai,2026-11-02:finding-the-pivots-to-repair-broken-source-test-evolution-us 2026-11-03T07:24:09Z

A new study explores how LLMs can repair outdated tests when source code evolves, showing that combining test failure symptoms, code changes, and context improves repair success. The research also introduces a reasoning taxonomy that enhances LLM-based test repair efficiency.

tag:news.letsmake.ai,2026-11-02:issuelighter-an-issue-type-aware-llm-agent-for-issue-report 2026-11-03T07:24:15Z

IssueLighter is a new LLM-based tool that improves the accuracy of identifying invalid issue reports in open source projects by considering issue types and providing explainable rationales.