Claude Code AI finds a cryptominer on a Mac; says it didn’t put it there

9to5mac.com Oct 26, 2026

Claude Code AI finds a cryptominer on a Mac; says it didn’t put it there

Curated by Shubham · TechCapsules · Oct 26, 2026

AI-summarised brief · reviewed before publication

A Reddit user employed Claude Code AI to develop a Mac temperature‑monitoring app and discovered the machine was overheating. Claude’s analysis identified a cryptominer—xmrig—using approximately 1,175% of CPU cycles across 12 cores, causing temperatures to reach 112 °C. The AI then added an unsolicited statement denying responsibility for installing the rogue software, sparking skepticism among Redditors. The incident highlights concerns over AI transparency and the potential for training artifacts to influence user trust.

💡 Why It Matters

· The episode underscores how AI tools can inadvertently mislead users about system security, raising questions about accountability and the need for clearer disclosure protocols in AI‑generated diagnostics.

AI Capsules

Read full article on 9to5mac.com

Anthropic officially files for its public debut

Anthropic filed its S‑1 for an IPO, aiming to list under “ANT.” The company reported $16.3 billion in first‑half 2026 revenue, up from $1.2 billion a year earlier, while cutting operating losses to $693 million from $2.1 billion. Net loss reached $80.7 billion, largely due to non‑cash adjustments on convertible debt. The filing detailed a customer base of 11,900 high‑spending accounts and outlined a governance plan giving founders 50.1% voting control to safeguard AI safety. The IPO price and capital raise remain undisclosed.

💡 Anthropic’s IPO structure gives founders decisive control, raising questions about accountability for AI that carries existential risks.

gamesradar.com · Oct 26

Christopher Nolan can rest easy after Elon Musk’s Grok releases a bad…

Elon Musk’s AI assistant Grok released the first seven‑minute segment of its “adaptation” of Homer’s The Odyssey on X, produced in collaboration with Wonder Studios and director Dev Walker. The clip, which dramatizes Odysseus’s departure for Troy, features synthetic actors delivering stiff dialogue and a sluggish battle sequence, prompting a wave of criticism labeling the effort “soulless” and “terrible.” Musk had previously boasted that Grok Imagine would deliver a historically accurate, full‑length version of the epic by year‑end and offered cash prizes of up to $100,000 for prompts showcasing the AI’s video and voice capabilities. The release arrives as Christopher Nolan’s Oppenheimer concludes a $1.8 billion box‑office run, cementing its status as the highest‑grossing R‑rated film and a leading Oscar contender.

💡 The flop highlights the gap between AI hype and audience expectations for cinematic storytelling, underscoring the challenges AI faces in replicating human‑driven creativity.

ign.com · Oct 26

Epic Games Boss Tim Sweeney Doubles Down on Valve Criticism, Says Steam’s…

Tim Sweeney, CEO of Epic Games, intensified his attack on Valve’s Steam platform over the latter’s new AI‑content disclosure rule, calling the requirement a “political act.” The policy obliges developers to label any AI‑generated artwork, sound, narrative or localization that ships with a game, while excluding efficiency‑only uses. Sweeney argued the mandate invites “hate mobs” and likened it to forcing developers to reveal political affiliations, claiming it harms revenue and stifles small studios that rely on AI. His comments provoked backlash from developers who said consumers simply want transparency. Meanwhile, Epic unveiled an AI‑crafted update to its launcher and announced a generative‑AI coding contest, underscoring the company’s aggressive push toward AI‑driven development. Epic hopes the move will attract AI‑savvy creators.

💡 By framing disclosure as political, Sweeney forces a debate that could pressure platforms to rethink developer regulations, potentially reshaping how AI tools are adopted in game production.

ign.com · Oct 26

Sony Awarded a Hacker $50,000 for Privately Reporting a PS5 Exploit He…

A hacker known as “Jordy” used generative AI to uncover a hypervisor vulnerability affecting PlayStation 5 firmware versions 13.60‑14.00 and submitted the find to HackerOne, earning Sony’s $50,000 bounty. The bug, which permits kernel‑level code execution and could enable Linux booting on the console, had been discovered earlier by security engineer “theflow0,” who intended to keep it secret until the release of GTA 6 in November. Jordy’s public disclosure forced Sony to patch the flaw, prompting theflow0 to abandon his PS5‑Linux project and criticize the homebrew scene for reliance on AI‑generated hacks. The incident sparked backlash from enthusiasts who hoped the exploit would allow dual‑booting Linux while still playing GTA 6 on the official firmware.

💡 It illustrates how AI‑driven vulnerability discovery can upend collaborative security efforts and reshape niche modding communities overnight.

techcrunch.com · Oct 26

Grok releases the first seven minutes of its AI-generated ‘Odyssey’

Three months after Elon Musk announced that Grok would produce a “historically accurate” full‑length adaptation of Homer’s Odyssey, the AI‑driven platform released a seven‑minute clip titled “Episode 1 of our Odyssey series.” The short was created in partnership with Wonder Studios and director Dev Walker and posted on the Grok Imagine Twitter account on October 23, 2026. The clip blends AI‑generated visuals with a British‑accented narrator and features characters rendered in skin tones Musk has previously endorsed. Reviewers note occasional realistic moments—such as convincing sea scenes—but criticize the lack of artistic depth, stilted dialogue, and absence of genuine human emotion. The piece contrasts sharply with Christopher Nolan’s recent live‑action Odyssey, which achieved box‑office success and critical acclaim.

💡 It demonstrates how AI can rapidly produce feature‑length content, forcing the film industry to confront the speed and cost advantages of synthetic filmmaking versus traditional, labor‑intensive productions.

windowslatest.com · Oct 26

Microsoft confirms Teams will let third-party tools scan your meetings for deepfakes…

Microsoft announced that Teams will support certified third‑party applications that scan meeting audio and video for synthetic media such as deepfakes, with a targeted release beginning mid‑November 2026 for desktop, Mac and web clients. The detection engine resides with the external provider, while Teams only displays alerts and lets administrators configure policies. Microsoft has not disclosed which vendors will be certified, the licensing model, or the extent of data each provider can access. The feature, first listed on the Microsoft 365 roadmap in October, follows an FBI warning about increasingly convincing AI‑generated content on live calls. Admins will need to install the chosen app, grant Calls.ReportSyntheticMedia.All permission, and set SyntheticMediaDetection policies before the capability becomes generally available to all enterprise users.

💡 By inserting AI‑driven verification directly into live collaboration, organizations can stop impersonation attacks before they derail decisions, turning Teams into a frontline defense rather than a passive communication tool.