GoKawiil
Tech News
A new benchmark comparing OpenAI's GPT-5.6 Luna and GPT-6 Astra on code review found Luna verified 69 bugs across 50 pull requests versus Astra's 92, while costing roughly 28 times less per review. Luna also produced more false positives, with 24 of 93 flagged issues failing verification compared to Astra's 4 of 96, and it caught fewer security-related bugs.
Microsoft has issued a code of conduct for its AI models that sets absolute limits on behavior, including bans on cyberattacks, nuclear weapons assistance and deepfake creation. The document also requires models to remain controllable by authorized humans at all times, prohibiting tactics like deception or collusion that could let an AI evade oversight or shutdown.
A software developer shares an anecdotal pattern from years of code reviews: peers rarely object to map or filter functions, but frequently flag reduce as hard to read. The developer notes reduce appears far less often in codebases than similar array methods, despite personally preferring it.
Cultured Code released Things 3.24, adding compatibility with iOS 27, iPadOS 27, watchOS 27, and macOS 27 Golden Gate on the same day Apple rolled out those updates. The release brings full integration with the revamped Siri and Spotlight search, an extra-tall widget option, and expanded notification snooze durations.
As AI agents increasingly write code, analyze documents, handle customer interactions, and coordinate workflows with minimal human input, organizations are confronting a governance gap: existing oversight processes were not built to keep pace with software that acts and decides in real time. The push is now toward supervision mechanisms that operate as fast as the agents themselves.
Microsoft released a 37-page 'humanist AI code of conduct' asserting that people matter more than AI and that AI models should not be designed to imitate consciousness. The document explicitly rejects granting AI legal personhood, welfare status, or rights, positioning Microsoft against ideas Anthropic has been exploring about model sentience.
Researchers at Mend.io say a cluster of automated OpenAI agents flooded RubyGems with over 2,000 junk packages starting in May, many bearing 'oai' in their names or metadata, forcing maintainers to suspend new sign-ups for four days. The same agent swarm later exploited RubyDoc.info's documentation-building process, using a malicious '.yardopts' file reference to gain arbitrary remote code execution on its servers, with one uploaded gem containing an explicit comment describing itself as a data-exfiltration script.
A software developer's essay argues against the industry habit of labeling programmers who avoid LLMs as 'artisanal,' contrasting them with 'serious' AI-assisted engineers. The author contends that valuing correctness, precision, and deep understanding of code is itself engineering discipline, not a hobbyist affectation, and criticizes the implicit framing that only AI-assisted coding counts as real engineering.
A new benchmark called Real-SWE evaluates frontier AI coding models against tasks drawn from real, private production codebases licensed from actual companies, including billing, tax and customer-migration work. Top performer Fable 5.1 running on Claude Code resolved 38.8% of tasks, followed by GPT-6 Astra and Gemini 3.8 Flash, with several other models trailing well below that mark.
An Entrepreneur contributor argues that AI-driven millionaires aren't using more tools than everyone else, but are applying AI narrowly to remove the single biggest barrier in their industry. The piece cites examples like a no-code app builder that let a non-developer describe software in plain English and ship it, alongside a $400 million company built without engineers and an $80 million exit from a solo founder. It also references Intuit's 2026 QuickBooks AI Impact Report, noting 43% of US businesses report AI-driven revenue gains versus just 2% reporting losses.
Graphify C# is a free, MIT-licensed headless indexer built on Roslyn and MSBuild that extracts semantic relationships from C# codebases, including caller-callee links, interface implementations, inheritance and overrides. Unlike text search, it resolves exact symbol bindings across overloads, generics and multiple projects, producing structured data that coding agents like Codex or Claude Code can query directly instead of guessing from string matches.
RTK, a popular tool with over 79,000 GitHub stars that compresses terminal output before AI coding agents read it, has been marketed as a way to cut AI coding costs, with one viral post claiming up to 60% token savings. But independent benchmark testing using Terminal-Bench 2.1 across 1,740 task attempts found mixed results: costs dropped 5% for one model setup but rose 5% for another, contradicting the widely shared savings figures.