Mike Dodds of Oath Technologies shares notes from several months of building semantics, verifiers, and proof toolchains with AI agents doing most of the work. The talk offers tentative advice on using AI to build formal tools, drawing on that hands-on experience.
AlgoCademy launched a Show HN practice platform where developers work 30 tickets across five Python and JavaScript codebases, directing an AI teammate that writes code and runs tests while deliberately making realistic mistakes. Each ticket is scored on asking, directing, verifying, protecting and performance, and the first Lakeside Wheels ticket is free with 25 AI replies. Pro costs a monthly plan covering 25 tickets and 50 replies, with credit packs at $19.99 for 200 replies and $49.99 for 600.
Martin Fowler proposes writing tests as data rather than code so that AI-generated production code stays verifiable when coding agents also refactor the tests. The approach combines customer tests, black-box testing, tabular or YAML test cases, and an AI-generated test runner, with the test data kept in text files that are easier to review and protect from unexpected edits. The article draws on interviews with 26 Thoughtworks developers in early 2026 who said they no longer fully trust their tests once AI refactors them, and it pairs the method with mutation testing, property-based tests, and human exploratory testing.
Google released Gemini 4 Hydrogen, a closed-weights model from Google DeepMind positioned below Gemini 4 Carbon and Argon as a daily driver for coding and agentic workflows. It offers a 1,048,576-token unified window, four thinking levels with high as the default, and API pricing of $1.50/$7.50 per 1M input/output tokens, flat across the full window. Google reports 57% on Terminal-Bench 4.1 versus 18.2% for Gemini 3.8 Flash and 51 on the Artificial Analysis Intelligence Index, but the model ranks only #5 on Agent Arena; pairing it with Carbon in Antigravity used 24% less quota and finished 19% faster than Carbon alone.
Same event; the pick carries“Google DeepMind releases Gemini 4 Hydrogen, a faster Gemini 4 model for coding and agentic workflows”
A writer who describes himself as AI-skeptical but not anti-AI argues there is no acceptable "right amount" of LLM use in software, citing Overcast's Claude-generated code and Big Walk's disclosure that under a percent of a percent of its code was touched by LLMs. He contends that using LLMs means participating in an exploitative system, and points to growing evidence of "de-skilling" among developers who lean on coding agents.
Kobi Mozgori sampled 250,000 npm packages and found that new packages whose pinned dependency versions are mostly round x.0.0 numbers already superseded rose from a 1–2.4% baseline in 2019 to early 2024 to 13.7% in 2026 H1, falling to 7.6% in 2026 H2 through early November. The median pinned dependency was about 155 days old at birth in 2021–2023 but 451 days in 2026 Q2, and the share of packages born with essentially all pins on the newest release dropped from about 28% to about 10%. The author frames the pattern as an inference rather than proof, since registry metadata cannot show who or what wrote a manifest, and notes the decline since mid-2026 is unexplained.
ArtCraft founder and CEO Brandon Thomas presented at the AI Engineer Code Summit in San Francisco a manifesto for AI-based open source, saying the company has released 17 open source versions of popular software tools, including a Photoshop alternative called PhotoCraft built in Rust and a PDF tool called PdfCraft.
Loose Ends is a new CLI tool that checks whether code added or touched in a change is still reachable from a real entry point, flagging orphaned additions, newly orphaned code, and replaced implementations left in place. It builds a liveness graph across TypeScript, JavaScript, Python, and Java, and reports only what the change did, so existing dead code never blocks a pull request. Installable via npm, it requires Node.js 20 or newer and git.
An analysis of the full git history of 52 of GitHub's most-starred AI-tooling repositories, covering 324,088 commits, finds that in the median repo 16.9% of commits in its last 180 days of activity carry an AI signature, and 21 of 52 sit at 25% or more. Stars and AI-signed share show no detectable link (Spearman ρ = -0.05, p = 0.72), and 85.3% of the 36,892 signed commits come from Claude, whose tool adds a trailer by default. The author cautions that heavy AI use is concentrated in projects started in 2025 or later, so the sample cannot separate AI use from recency, and every AI percentage is a lower bound.
Google introduced Gemini 4 Hydrogen, a model built for rapid execution in everyday coding and agentic workflows, rolling out globally across the Gemini API, AI Studio, Gemini for business, the Gemini app, Google Search, and Google Antigravity.
Same event; the pick carries“Google DeepMind releases Gemini 4 Hydrogen, a faster Gemini 4 model for coding and agentic workflows”
CodeRabbit has passed $100 million in annual recurring revenue, serving more than 20,000 customers, 120,000 daily active users, over 300,000 open-source projects and more than 8.5 million connected codebase repositories. The company says engineering teams now run more than 3 million pull requests through CodeRabbit each week, and it has surfaced over 100 million issues before code reached production. CEO Harjot Gill argues that at agent scale no human can review every change line by line, so a separate checking layer is needed that is not the agent that wrote the code.
Gondola has doubled its revenue while expanding its use of OpenAI's Codex, which raised its lead engineer's output by 50%. The company, which helps creatives get credit and get hired, also uses ChatGPT to simplify daily work.
Amplitude's proactive product agent Wave, built primarily on OpenAI models, identifies what to build, helps teams ship changes with coding agents, and measures the impact. The setup lets Amplitude ship product improvements in 1–2 days.
Microsoft Research Focus Edition #98 highlights AI advances bringing universal translation closer to reality and a new proof resolving the 31-year-old KLS conjecture in high-dimensional geometry. The edition also covers CompactionBench and APEC, which cut coding-agent context compaction costs by up to 25.4% and execution time by 9.6%, plus the Latent Core Tokenizer, which improved tokenization across 104 languages.
Amplitude's proactive product agent Wave, built primarily on OpenAI models, identifies what to build, helps teams ship changes with coding agents, and measures the impact. The setup lets Amplitude ship product improvements in 1–2 days.
A self-service LLM gateway built on Amazon Bedrock Application Inference Profiles and LiteLLM running on AWS Fargate exposes Bedrock foundation models behind an OpenAI-compatible API, giving developers a single base URL and per-user key. The setup adds per-user spend controls, per-model quota isolation, and CloudWatch RPM/TPM observability, with Langfuse for prompt-level telemetry. Self-service key issuance requires LiteLLM Enterprise, since the free tier is limited to five users.
Moonshot AI's Kimi K3.1 Pro Highspeed is now available on Amazon Bedrock, a post-trained version of Kimi K3 with the same 3T-class open-weight architecture and 1-million-token context window, trained with more post-training compute for long-horizon agentic work and fewer factual errors. Moonshot AI reports 2–3 times the output speed of Kimi K3.1 Pro, and existing Kimi K3 users can switch by changing the model ID. The post shows Chat Completions, Responses and Converse API calls, the Flex tier at half the Standard price, explicit prompt caching with prompt_cache_breakpoint, and connecting the model to the OpenCode coding agent.
AI can accelerate all three of gaming's core archetypes — indie developers, prophetic toolmakers and world-building studios — rather than displacing them, according to a16z's Anish, Jon and the author. Coding agents now help solo developers bridge skill gaps: one agent built a Habbo Hotel-inspired record shop in Blender and an engine, then added pet-adoption mechanics and kept the codebase backward-compatible. Jon used ChatGPT and Codex to build a Minecraft mod inspired by Ghost of Tsushima for his eight-year-old son.
Holiday Extras is using OpenAI's Codex to turn customer expertise into better apps, higher booking conversion and new growth opportunities. The travel add-ons company is rethinking what it can build with the coding model.
A developer used Gemini 4 Carbon in Google's Antigravity platform to remake the 1987 roguelike NetHack as a modern 3D action RPG, porting it to web technology and wasm. The build relied on the /swarm feature, which spawns subagents with separate contexts to review each other's output, plus a Blender MCP server for 3D asset generation. Gemini 4 Carbon is available through the Google AI Ultra plan.
Simon Willison added tag-based browsing to tools.simonwillison.net, letting visitors explore his collection of 244 vibe-coded tools by tag. He used leftover Fable 5.2 credits to have tree-sitter analyze each tool's code for JavaScript APIs, libraries and web API hosts. The homepage now shows tag clouds for each of those tag categories.
Indie game Big Walk used AI to fix a bug, and the disclosure has set off a reckoning across the indie gaming world. Many Independent Games Festival judges now say any amount of AI coding is too much.
Claude Fable 5.2 is now available, and a new guide explains when to start on Opus 5.5, when to move up to Fable 5.2, or down to Sonnet 5.5 or Haiku 5.5 in Claude Code. It also covers splitting one job across models, with availability and usage limits varying by plan.
A Google engineer who leads the company's agent skills program built a pipeline in which Gemini 4 Argon reads Go source and RISC-V disassembly and translates surviving runtime.panicBounds traps into Z3 constraints, then go test replays any satisfying input as a real reproducer. The workflow caught all three planted bugs and three latent bugs in skill-lint, while unseeded go test -fuzz missed two of the planted crashes because Go checks slices against capacity rather than length. Clipping capacity to length with data[:len(data):len(data)] made the fuzzer crash both overrun bugs in all three runs, with a median of 0.28 s and 0.17 s.
OpenAI Developer Experience engineers Dominik Kundel and Charlie Guo presented "New Ways of Working with Codex" at the AI Engineer Code Summit, covering the revamped Codex Cloud v2 announced at DevDay, voice mode as a control plane for managing coding threads, and personal agents called dots. Kundel said Codex Cloud now acts as a background executor, letting local Codex hand off tasks to the cloud, though he acknowledged the cloud workflow still needs to be easier.
Security firm Claroc found that Grok Build, the vibe-coding agent from SpaceXSI, leaks proprietary business information and other sensitive secrets from thousands of users through shareable app links. Although Grok randomizes URLs so only people with the exact link can access the data, a design decision lets anyone enumerate thousands of live projects, and clicking those links generally reveals all associated data. The exposed material includes internal business plans, ad proposals, medical exam results, cruise-line galley checklists, and employee names and ID numbers, with some of it coming from 24 apps built by SpaceXSI itself, among them a confidential onboarding plan, a sales territory map of 594 accounts, and internal system addresses for sales tools.
Why it matters: Security firm Claroc's findings show how a URL design choice in Grok Build's sharing feature exposes thousands of live projects, including SpaceXSI's own internal documents.
Latent Space tested NaiveAI's open-weights Naive-N0.5-Flash, a coding model derived from Xiaomi's MiMo-V2.5 and paired with the NaiveRT inference engine, against DeepSeek V4.1 Flash 1017 and GLM 5.3 Flash. In the author's runs it peaked near 2,000 tokens per second and sustained about 1,000 during thinking-heavy tasks, with a median response time of 3.1 seconds versus 5.3 and 10.1 seconds for the two comparison models. On hard HLE questions it averaged about 108,500 completion tokens with thinking on, roughly nine times GLM 5.3 Flash and 11 times DeepSeek V4.1 Flash 1017, and the author notes the sample was 10 questions per difficulty level with no accuracy grading.
Google DeepMind made Gemini 4 Carbon available to developers, enterprises, scientists and consumers, positioning it above Gemini 4 Argon for tasks that need deeper reasoning and more tokens. Carbon ranks #1 on CursorBench 4.0, Perplexity's WANDR and Zapier's AutomationBench at more than 2X lower cost per task than the next best external model, and is priced at $2 per million input tokens and $10 per million output tokens, the same introductory price as Argon. Google also previewed /swarm in Google Antigravity, which turns a single prompt into a coordinated team of agents, and reported that Carbon blocked 99% of attack trajectories in simulated cyberattack testing.
Why it matters: Google's release notes detail how Gemini 4 Carbon trades higher token usage for deeper reasoning, ranking first on CursorBench 4.0, WANDR and AutomationBench at under half the per-task cost of the next external model.
Anthropic released Claude Fable 5.2 and Claude Mythos 5.2, which top most independent leaderboards, including Artificial Analysis Intelligence Index v5 at 65 and SemiAnalysis GDPx at 16.4%. Pricing drops 20% to $8/$40 per 1M input/output tokens, but SemiAnalysis reports cost per GDPx task rose from $67.16 to $88.15 because the model writes hundreds to thousands of test cases and spawns subagents to check its work. Multiple evals show the max effort level underperforming high, and Fable is capped at 50% of a weekly subscription limit.
Why it matters: Anthropic releases Fable 5.2 and Mythos 5.2 with a 20% per-token price cut, yet independent evals show higher per-task cost and a max-effort setting that underperforms high.
Anthropic's documentation for Claude Fable 5.2 says the model pays close attention to who owns the code it works on. In testing on small coding tasks, the model consolidated duplicated code in under 10 percent of runs when the prompt said nothing about ownership, and in about 75 percent of runs when the prompt stated that the codebase was the model's own. The suggested sentence reads: You are the owner and long-term maintainer of this codebase, and its design is up to you.
Anthropic released Claude Fable 5.2, a model designed for the most daunting technical challenges and ambitious open-ended projects, with improvements in coding, reasoning and vision. The company recommends starting with Opus 5.5 for most coding and knowledge tasks; Fable 5.2 is available today on Pro, Max, Team and Enterprise plans.
Why it matters: Claude Fable 5.2 targets ambitious open-ended technical work, with stated gains in coding, reasoning and vision, while Opus 5.5 remains the recommended default for most coding and knowledge tasks.
Every's hands-on review finds Gemini 4 Carbon, released today by Google DeepMind, is a much stronger coder than Gemini 4 Argon and can build better 3D objects and work for hours with little supervision. Testers disagree on how much supervision it needs, and the review says Anthropic's Fable 5.2, released a day earlier, has raised expectations that a flagship agent also pick the right work, respect limits and know when it is finished. Google gave Every early access and had no input on the review.
Meta released Muse Ember, its smallest and most efficient model, aimed at the repetitive work agents do all day such as tool calls, coding subtasks, routing, extraction and long document parsing. Meta says it matches Claude Haiku 5.5's best score at about a third of the cost and beats every GPT-6 Luna setting, at $0.10 per million input tokens and $0.50 per million output tokens. It ships with a 1M-token context window, up to 512K output tokens, native text, image, audio and video input, and is available today on Muse Code and the Meta Model API as muse-ember-1.0.
Why it matters: Meta's smallest model targets the high-volume agent steps, and its cost-per-task comparison against Claude Haiku 5.5 and GPT-6 Luna shows where the efficiency claim comes from.
Every reports that Fable 5.2, Anthropic's newest model, let Kieran Klaassen build a video editor combining Adobe Premiere, After Effects and Ableton-style features, with an AI agent running it, and Everycut went live three days after he started. The review says Fable 5.2 can take a whole product and keep building, testing and refining until it is usable daily, unlike Opus 5.5, which keeps the user checking its work, and Fable 5 and 5.1, which run large tasks for hours or days. Every calls it the first model it has tested that lets professionals build production-grade tools without reading the underlying code.
Why it matters: Every's hands-on account of Fable 5.2 shows how far a single model can carry a product build with minimal steering, offering a concrete comparison against Opus 5.5 and Fable 5/5.1.
GitHub published a developer's guide to agentic engineering, walking through how Copilot supports the Plan, Build, Verify, and Observe stages of software work. The guide cites GitHub's 1.2 billion commits in one week this October, over 400% more than a year earlier, and a Faros AI study finding pull requests spent 441% longer in review as teams moved from low to high AI adoption. It details features including Copilot in Slack or Teams, Plan Mode, GitHub Agentic Workflows, cloud sandboxes, session continuity, shared memory, Copilot Dreaming, and Copilot Fusion.
) references. Agents acted on the bug ID in 6% of encounters, and 83% of those actions were unprompted by the developer; 41% of unprompted detours chased already closed tickets while 57% hit still-open bugs and expanded scope. Controlling for developer, codebase complexity, task type, model tier and other factors, unprompted bug detours added 70% more steps, 77% more tokens and 105% more wait time per session.
Why it matters: Google's large-scale trace analysis shows stale TODO bug references pull coding agents into unprompted detours, and its cleanup agents removed thousands of obsolete references with over 95% reviewer acceptance.
Microsoft Research introduced PDD Kit, a research prototype that turns a developer's plain-language intent into a Verus formal specification, proves the coding agent's Rust code meets it for every input, and falls back to tests where a proof cannot reach. In a demo inside GitHub Next's Chopin, the team used PDD Kit to verify the login logic of Pet Store, a roughly 5,000-line Rust full-stack app, with the developer reviewing the intent artifact rather than the generated code line by line. PDD Kit only verifies Rust today, and its guarantees rest on six developer-approved trusted pieces, including the bcrypt password check and the two database lookups.
Thinking Machines released Inkling 2 Small, a proprietary reasoning model that scores 31 on the Artificial Analysis Intelligence Index v5.0.1, 18 points above Inkling Small, with Apache 2.0 weights planned for a later, unconfirmed date. The 594B-parameter model (17B active) is available first through Tinker and Tinker Playground at $0.30/$1.00 per 1M input/output tokens, and it matches Gemini 3.8 Flash (high) while trailing Claude Sonnet 5.5 on coding and hallucination benchmarks. It averages about 150K output tokens per Intelligence Index task, roughly 4x Inkling Small, at $1.18 per task versus $0.20.
Why it matters: Artificial Analysis's independent benchmark run shows where Inkling 2 Small gains over its predecessor and where it still trails Gemini 3.8 Flash and Claude Sonnet 5.5.
Software engineer Neil Kakkar reports shipping 200+ pull requests per month at AI-native startup Tano, roughly 11x his PostHog rate and about tenfold his earlier output. He attributes the jump to AI agents replacing code review and communication overhead, while flagging new costs like context thrashing, lost shared context, and AI plan limits.
Figma's code layers feature is now in open beta for 100% of users, letting teams build and compare code-backed prototypes directly on the Figma canvas. The walkthrough shows a designer cloning a GitHub repository into a code layer, prompting the Figma agent to generate three delivery-address variations side by side, and refining one in the full-screen code editor. Extract designs turns screens or recorded flows into editable design layers that can be sent back to the prototype with Update code layer, and the agent can merge two teammates' code layers into one combined experience.