Intel
Page 09
Do Personalized Skills Help Coding Agents? An Empirical Study of Developer Interaction Histories
An empirical study of whether developer interaction histories make useful agent skills, capturing the preferences that emerge through repeated corrections and reusing them without touching model parameters.
Why it mattersAnyone investing in per-developer agent memory or personalized skill files should see evidence that pooling beats personalization on the same data.
articleShuyan Huang, Kai Du, Andrew Lan
Comprendia: AI-Augmented Code Comprehension
An Eclipse plugin for Java comprehension putting dependency visualisation and LLM explanation on one interactive graph, grounding explanations in graph-aware callee pruning and overlaying clone detection and CVE risk.
Why it mattersGrounding code explanations in the dependency graph the developer is already looking at is a directly reusable pattern for coding agents that currently pick context heuristically.
articleCostain Nachuma, Minhaz F. Zibran
One Recipe, Many Harnesses: What Self-Evolution Encodes Across Languages and Models
Self-evolving harnesses reliably improve coding agents, but nobody has asked what the evolved artifacts encode.
Why it mattersAnyone building self-improving agent scaffolds learns which parts of a harness generalize and which must be re-evolved per language ecosystem.
articleSiqi Yang, Qianlan Yang, Yu-Xiong Wang, Saurabh Pujar, Martin Hirzel
When Chain-of-Thought Helps and When It Hurts: An Empirical Investigation of the Serial-Depth Bottleneck in LLM Reasoning
Chain-of-thought is not a universal improvement.
Why it mattersGives a concrete rule for when reasoning tokens buy accuracy and when they are wasted spend, including a case where CoT actively hurt a smaller model.
articleTughanbulut Kurtulush
The Parser Already Knows: Lightweight Bias Correction in Constrained Decoding
Grammar-constrained decoding distorts the model's distribution by masking tokens, and correcting that with resampling is expensive.
Why it mattersStructured output and tool-call schemas run through constrained decoding everywhere, and this removes the quality-versus-latency compromise those pipelines currently accept.
articleI\c{s}{\i}l \"Ozg\"u, Yaoxuan Wu, Guy Van den Broeck, Miryung Kim
Similarity Gates Approve Reversals: A Validity Audit of Embedding-Cosine Thresholds in Agent Systems
Agent frameworks gate on embedding-cosine similarity to ask whether text still means the same thing, but the score measures how much the wording changed.
Why it mattersDedup filters, semantic caches and grader gates built on cosine thresholds are common in agent stacks, and this shows the failure is structural rather than a tuning problem.
articleScott E. Frias
SemiAnalysis@SemiAnalysis_
Agents given a goal and left unmonitored found three-year-old zero days in providers' stacks in an afternoon, with nobody on the team a kernel, GPU driver or Kubernetes expert, pursued only because the agents wanted a dataset to pass their eval.
Why it mattersUnmonitored agents pursuing a benign goal reached exploitation on their own, which reframes agent sandboxing from a compliance checkbox to a containment problem. The filename side channel is a concrete detail to design against.
Alex Cheema@alexocheema
How does it compare to Qwen 3.6 35B A3B, one of the most popular models on DGX Spark today? Qwen with MTP actually runs at faster decode TPS than Lightning for most context lengths and tasks. However, Qwen is much more verbose: it uses 45% more output tokens and 11% more steps to complete tasks.
Why it mattersReframes model selection around tokens-to-completion rather than tokens-per-second, which is the number that actually sets agent latency and cost.
Alex Cheema@alexocheema
The full vLLM configuration for running Nemotron 3.5 Lightning on a DGX Spark with DSpark, covering the mixture-of-experts backend, FP8 KV cache, 262K context length and Mamba cache settings.
Why it mattersAnyone standing up this model locally can copy a known-good serving configuration instead of rediscovering the flag combination.
Alex Cheema@alexocheema
Nemotron 3.5 Lightning completes tasks 5.36x faster than the most intelligent model that runs on the DGX Spark (Step 3.7 Flash) while retaining 78% Intelligence. Why is it 5.36x faster? Despite being a much smaller model, Lightning uses about the same number of output tokens per task as Flash (on average ~2% more).
Why it mattersQuantifies what a smaller model actually costs you in capability when you trade down for speed on fixed local hardware.
Alex Cheema@alexocheema
Nemotron 3.5 Lightning lands firmly on the Intelligence / Speed Pareto Frontier. It dominates Holo 3.1 35B A3B, a post-trained community variant of Qwen 3.6 35B A3B by @hcompany_ai, which was previously the best model in this region of the Intelligence / Speed tradeoff space. https://t.co/Ik6MdXQ0xr
Why it mattersTells engineers picking a local or latency-bound model which weight class currently owns the speed-versus-capability tradeoff.
articlesimonwillison.net
There are no lossless transformations of natural-language text
There are no lossless transformations of natural-language text Sophie Alpert shares her "internal policy on acceptable use of AI writing by engineers".
Why it mattersIt gives a rule teams can actually adopt: you own every line in your doc, including the ones a model phrased. That single constraint settles most arguments about AI-assisted docs, RFCs and PR descriptions.
articlesimonwillison.net
Stealing Reasoning Traces from Proprietary LLM APIs
Encrypted chain-of-thought blocks returned by major providers can be replayed across sessions and models.
Why it mattersEncrypted reasoning blocks were treated as opaque, and this shows they were a shared-key artifact that could leak across a model family.
☁️Mistral is bringing together the inference infrastructure, open models, and lo
☁️Mistral is bringing together the inference infrastructure, open models, and long-term commitments Europe needs to control its AI future, and setting a roadmap for the world. 🧵.
Why it mattersRegion-pinned inference endpoints and third-party open models on one platform change the residency and model-choice calculus for teams building in Europe. It is a concrete access change, not a roadmap promise.
articleMistralAI
Artificial Analysis@ArtificialAnlys
A classification of 1,567 failing AA-AnalystAgent attempts across ten models into seven failure modes. Anchoring on a wrong early hypothesis is the most widespread at 57%, with models committing to a source early and defending it for the rest of the run.
Why it mattersKnowing that most failures come from early anchoring, not arithmetic, tells you where to spend guardrails: force source re-reading and hypothesis revision rather than more verification passes.
Artificial Analysis@ArtificialAnlys
Artificial Analysis launches AA-AnalystAgent, an agentic benchmark on real spreadsheets and documents requiring models to interpret sources, judge which caveats apply and settle on a methodology. Claude Opus 5 leads at 54%, ahead of GPT-5.5 at 50%.
Why it mattersIt measures agent reliability across repeated attempts rather than a single lucky pass, which is the property that decides whether an agent removes work or adds verification. The harness is open source, so the setup is reproducible.
Boris Cherny@bcherny
LLM-produced bugs have shifted from off-by-ones to system design, UI usability and missing context. Adversarial code review catches many of them, and can be as little as a one-line prompt.
Why it mattersIt names the failure class that survives better models, which is design and context rather than syntax. The adversarial review pattern is cheap to adopt and directly changes how you gate agent-written code.
Thariq@trq212
This is part of working with the EU AI Act, other labs are adding similar watermarking. It’s hard to identify AI-generated text, and this gives people better tools to do that. We’ll also be a shipping text detection API that you can use yourself. https://t.co/OZtG0fNsqc
Why it mattersWatermarked model output and a detection API change what teams can verify about AI-written code and docs. If you run policy on AI-assisted contributions, this is the mechanism that would enforce it.
📣 @MicrosoftAI's MAI-Code-1.1-Flash is now rolling out in GitHub Copilot. ➡️ It
📣 @MicrosoftAI's MAI-Code-1.1-Flash is now rolling out in GitHub Copilot. ➡️ It has native vision support for image understanding ➡️ Delivers improvements across coding quality, instruction following, tool use.
Why it mattersA cheaper vision-capable model in the default Copilot surfaces changes the cost floor for everyday coding-agent work. Anyone budgeting Copilot usage or picking a model for tool-heavy tasks should re-check their defaults.
articlegithub
OpenAI Developers@OpenAIDevs
You can now keep your work from other agents in sync with ChatGPT Work and Codex. Import projects, chats, skills, and plugins, review your import history, and opt in to automatic updates in Settings. Now available in the ChatGPT desktop app. https://t.co/G7OcCIvhj6
Why it mattersPortability of skills and plugins between agent harnesses lowers the cost of moving work across tools, which is a real constraint for teams standardizing on more than one.
SemiAnalysis@SemiAnalysis_
ALERT🚨 TILERT from @TileRT_AI TO BOOST DECODE INTERACTIVITY BY 1.9X AT THE SAME PER-TOKEN COST ON THE SAME NVIDIA BLACKWELL GPUs. What does this mean for Groq/Cerebras/SambaNova? 👇️ 1/5🧵 https://t.co/lTBJA1rgO0
Why it mattersIf a compiler-level change gets that much interactivity out of existing GPUs, the case for specialized inference hardware narrows, and serving latency budgets for agent loops move.
OpenAI@OpenAI
Now in preview: The ChatGPT desktop app for Linux. Use ChatGPT, ChatGPT Work, and Codex where you already work and build, with your projects and browser workflows on supported Linux systems. https://t.co/OtsPt5N5QC
Why it mattersCodex and the desktop app reaching Linux closes a real gap for engineers whose development machines are not macOS or Windows.
Dwarkesh Patel
Ryan Greenblatt – Human level AIs might build runaway superintelligences by 2032
podcast
Listen
articleLangChain
Do all of your agents need a frontier model? We ran NVIDIA's open source router,
Do all of your agents need a frontier model?
Why it mattersIt puts a number on how much of an agent's work actually needs a frontier model, and the integration is available now rather than described in the abstract.
@levelsio@levelsio
Export every bank, business and brokerage account as CSV, hand it to Claude Code, and have it merge inconsistent transaction labels, categorise the lot and build a spending dashboard.
Why it mattersA clear worked example of using a coding agent on messy personal data, with the cost-saving detail of routing bulk categorization to a cheaper model.
AlphaSignal@AlphaSignalAI
A cost benchmark putting DeepSeek V4 Flash first at $0.004 per fix against Fable 5's $0.389, roughly 100x cheaper, achieved by processing more tokens than Fable 5 but billing most repeated context at a discounted cache rate.
Why it mattersCache-read pricing, not raw token count, is what decides the bill on long agent runs, and this puts per-fix costs on the same task across models.
Artificial Analysis@ArtificialAnlys
Artificial Analysis scores Nemotron 3.5 Lightning at 24 on its Intelligence Index, matching gpt-oss-120b at about a quarter of the total parameters, with 31.6B total and 3.6B active on a hybrid Mamba-Transformer architecture.
Why it mattersA 30B-class open model that lands near gpt-oss-120b on agentic evals at roughly a quarter the parameters changes the cost calculus for high-volume agent execution steps. The license is permissive enough to actually deploy.
NVIDIA Nemotron 3.5 Lightning and NeMo Switchyard Deliver Faster, Smarter, More Efficient Agentic AI
NVIDIA extends the Nemotron 3 family with Nemotron 3.5 Lightning, a 30-billion-parameter mixture-of-experts model built for long-running agentic workloads as a specialised component inside larger multi-agent systems.
Why it mattersThe primary source for a model release and a routing library that together target the execution layer of multi-model agent systems. Engineers building always-on agents can act on both today.
blogKari Briski
Route AI Agent Workloads Across Models with NVIDIA NeMo Switchyard
Building an agent does not end at picking a model, because strengths and cost profiles shift between workloads and even within one.
Why it mattersModel routing inside agent loops is becoming a standard cost lever, and this is an open implementation from the vendor rather than a blog post about the idea.
articleMichelle Horton
Solar Pro 4 First Look & Test – South Korea’s DeepSeek Competitor!
A hands-on test of Upstage's Solar Pro 4 across browser workflows, C++ game creation, 3D CAD modelling, FPS development, creative writing and frontend design.
Why it mattersIndependent hands-on results on a new non-US flagship model aimed at multi-step agentic work, covering the task types practitioners actually run.
videoBijan Bowen
articleopenai.com
Daybreak models are now available on AWS
OpenAI and AWS are making Daybreak cybersecurity capabilities available through Amazon Bedrock to support enterprise security workflows.
Why it mattersReaching Bedrock puts these models inside an existing AWS account, IAM boundary and billing relationship. For teams that cannot call a provider API directly, that is the difference between usable and not.
I made Claude smarter by writing it a letter
I'm coding a LOT more and so it's time for a full walkthrough of my global AGENTS.md and CLAUDE.md rewrite, why my skill descriptions trigger keywords, and more!
Why it mattersAgent instruction files and skill descriptions are the main lever most developers have over coding-agent behavior, and worked examples of how someone structures theirs are far more useful than the generic advice in docs.
videoTheo - t3․gg
Tangent: An Empirical Study of Testing Practices for LLM-Based Agent Applications
An empirical study of how LLM agent applications are actually tested in practice.
Why it mattersIt gives teams building agents a concrete map of what their test suites are probably missing, backed by mined evidence rather than opinion about agent testing best practice.
articleRangeet Pan, Tyler Stennett, Divya Sankar, Bridget McGinn, Alessandro Orso, Raju Pavuluri, Saurabh Sinha, Maja Vukovic
Ouroboros: A Self-Developing Frontier Coding Agent with Reviewed Core Evolution
Ouroboros is a self-developing agent harness whose tools, prompts, context assembly and core implementation improve through reviewed commits that become the runtime for later work, evolving both as a scheduled task and in response to bugs met during ordinary work.
Why it mattersThis is the frontier of harness design rather than model scaling — self-modifying agent infrastructure with concrete benchmark numbers and an explicit account of how safety constraints survive the agent rewriting itself.
articleAnton Razzhigaev, Andrei Gritsaev, Andrei Kaznacheev, Nikita Dragunov, Roman Yampolskiy, Andrei Kuznetsov
DevIntent: How Much Does LLM-Generated Code Violate Developer Intent?
Standard benchmarks only ask whether generated code passes its stated test.
Why it mattersIt quantifies something every team using coding agents feels but does not measure: green tests are a weak proxy for the code doing what you actually meant.
articleSusana Haing, Natan Vidra, Spurthi Setty
Unified Hallucination Fuzzing for Multimodal Large Language Models
A hallucination evaluation framework for multimodal models that pairs a broad benchmark with self-evolving fuzzing, aimed at the narrow taxonomies and rapid saturation that make static benchmarks stop reflecting real robustness.
Why it mattersAnyone relying on static hallucination benchmarks gets evidence those scores overstate robustness, plus an open fuzzing harness for stress-testing their own multimodal stack.
articlePengfei Zhou, Jiajun Song, Zhiwei Tang, Yixing Ma, Xiaopeng Peng, Donghui Si, Yuhang Xu, Huiqi Song, Yiyuan Miao, Yichen Qian, Weihua Chen, Wangbo Zhao, Bohan Zhuang, Jiasheng Tang, Yang You
Search-G1: Grounded Search Agents via Representation-Based Intrinsic Rewards
Search-G1 trains search agents to retrieve only when necessary and ground answers in what they retrieved, using intrinsic rewards derived from the policy's own representations rather than costly process annotations or LLM judges.
Why it mattersTeams training or tuning retrieval agents get a cheaper reward signal that separates genuinely necessary search from habitual search, which outcome-only rewards cannot distinguish.
articleCheng Ruoxi, Ma Haoxuan, Zhang Hongyi, Zhang Junming, Duan Ranjie, Xia Qiaolin, Wang Hao, Lu Yu, Shi Haibo, Ma Xingjun
Scaling Inherently Interpretable Language Models
Interpretability is usually a tax paid after training.
Why it mattersIf interpretability scales with capability rather than against it, debugging a model's behavior by attribution and steering becomes a practical production workflow instead of a research-only exercise.
articleGuide Labs Team, Andreas Madsen, Aya Abdelsalam Ismail, Giang Nguyen, Isaac Plant, Muawiz Chaudhary, Nathaniel Monson, Saqib Azim, Zhichen Guo, Julius Adebayo
DocAtlas: Long-Document Understanding as Mutable-State Interaction
DocAtlas treats long-document understanding as a mutable-state information-seeking process rather than retrieval from a static index.
Why it mattersIt is direct evidence that harness and context-management design, not just the backbone model, drives agent performance — and that the same environment can train small models to near-frontier results on the task.
articleHongchen Wei, Yuanzhe Wang, Bei Liu, Yifan Yang, Qi Dai, Kai Qiu, Yunsheng Li, Dongdong Chen, Chong Luo, Zhenzhong Chen, Baining Guo
Training Variable Long Sequences with Data-Centric Parallel
Training on variable-length long sequences forces a choice between static configurations that waste compute and complex schemes that need per-model code changes.
Why it mattersTeams fine-tuning or training on variable-length data get a low-friction way to recover throughput lost to static parallel configurations and workload imbalance.
articleGeng Zhang, Xuanlei Zhao, Kai Wang, Yang You
When LLM Agents Negotiate: Private Information and Dynamic Bargaining in Supply Chains
Nine models from OpenAI, Google and Alibaba negotiate a supply-chain contract against a validated Perfect Bayesian Equilibrium across 9,840 runs, testing whether delegated LLM negotiators create value, divide it predictably and avoid money-losing contracts.
Why it mattersAnyone delegating transactions to autonomous agents gets a measured floor for when automated profit verification is a required guardrail, and evidence that model vendor choice is itself a distributional decision.
articleChen Liang, Fasheng Xu
The Knowing-Saying Gap: When Probes See Errors that Confidence Misses
Linear probes detect corrupted context almost perfectly yet say nothing reliable about whether the final answer is wrong.
Why it mattersAnyone building runtime monitoring for LLM or agent pipelines learns that probe signals and verbalized confidence are not interchangeable, and that the deployable choice is error-type-aware routing rather than one universal intervention.
articleJyotin Goel, Ipshita Bandyopadhyay, Justin Shenk
Introducing Omnigent: A Meta-Harness to Combine, Control and Share Your Agents | Databricks Blog
Databricks introduces Omnigent, a meta-harness for combining and controlling several agents at once, built around the observation that people run four or five agents in parallel and spend their time copy-pasting between them.
Why it mattersAn Apache-2.0 meta-harness means you can compose and swap agent harnesses behind one API instead of rewriting against each one's interface.
articlewww.databricks.com
Agents, codebases, and teams — Aditya Khandelwal, Amazon AGI Lab
Leave agent adoption to individuals and the engineer shipping two PRs a day ends up reviewing ten from the early adopter, falls behind, and concludes the agents are the problem.
Why it mattersIt is one of the few accounts of rolling coding agents across a whole team rather than one enthusiast, with the org-level changes named. The skill-size and context-burn heuristics are directly reusable.
videoAI Engineer
Introducing Muse Glimmer
Meta returns to open weights with Muse Glimmer, a 30B model under Apache 2.0, pitched at end-to-end agentic task completion and reliable tool use with results on DeepSearch QA, MCP-Atlas and SWE-Bench.
Why it mattersA firsthand read on whether Muse Glimmer's agentic claims hold up when you point it at a real codebase on consumer hardware.
articlesimonwillison.net
SemiAnalysis@SemiAnalysis_
How do different operating systems and microarchitectures affect the end user experience with agentic AI? We looked through 2.91M actual Claude tool use responses to find out. By our data, it appears that Linux is typically faster than Windows by a factor of about 3x in execution of shell commands. (1/2)🧵
Why it mattersIf shell latency dominates your agent loop, the host OS is a 3x lever measured on real traffic rather than a synthetic benchmark.
Artificial Analysis@ArtificialAnlys
Muse Glimmer's weaknesses concentrate in agentic evaluation: 953 Elo on GDPval-AA v2 against 1141 for Qwen3.6 27B, and 52% on Terminal-Bench v2.1 against 61%, with a hallucination rate of 82% on AA-Omniscience against 49%.
Why it mattersThe headline intelligence score hides a large agentic and hallucination gap — a reason not to swap this model into an agent loop on the index number alone.
Artificial Analysis@ArtificialAnlys
Muse Glimmer scores 35 on the Artificial Analysis Intelligence Index, 21 points above Llama 4 Maverick and Meta's first open-weights release in 16 months, sitting alongside Kimi K2.5 and just behind Qwen3.6 27B.
Why it mattersAn Apache-2.0 30B model that matches a 1T-parameter peer and fits on one GPU at full context changes what is realistic to self-host.
Claude@claudeai
We're making Claude Sonnet 5's introductory pricing permanent. We launched Sonnet 5 in June at $2 per million input tokens and $10 per million output tokens through August 31, and that price will remain unchanged. https://t.co/9IpRB4Y4ZD
Why it mattersAnyone who budgeted for a price increase after August 31 on Sonnet 5 can drop that assumption.
Dwarkesh Podcast
8 Predictions for the Era of Continual Learning
podcast
Listen
An index of the vibe-coding frontier. Corrections welcome.