Hyperscale Data Pivots to AI Data Centers, Allocates $100M to Bitcoin Treasury ↗
Hyperscale Data will allocate $100M to a Bitcoin treasury as it pivots into AI data centers and digital assets. Funding comes from selling its Montana data center and an at‑the‑market equity program. Its Michigan campus (now ~30 MW) aims for 70 MW in 20 months via on‑site natural gas and could expand to ~340 MW. Through Sentinum, the firm will continue mining and will publish weekly crypto holdings. Ault Capital Group divestiture is planned for Q1 2026.
AI Housing Approvals Could Repeat Robodebt Mistakes, Experts Warn ↗
NSW and federal politicians are pitching AI to speed housing approvals, calling it a 'gamechanger', but experts warn it risks sidelining planners' judgement, obscuring errors and nudging decisions in unseen ways. Lessons from the Robodebt debacle show automation can do harm if rushed. Nabavi argues governments should first define the problem, address root causes (labour, finance, social housing) and adopt responsible, participatory AI design rather than pursue quick fixes.
Senior Devs Scrub Hallucinations and Security Flaws from AI-Generated Code ↗
Vibe coding—using AI to write code—has pushed senior developers into roles of verifier and fixer, chasing hallucinated packages, deleted data and new security holes. A Fastly survey found 95% of devs spend extra time fixing AI-generated code, though senior engineers still often ship it. Developers say AI speeds prototyping but requires mandatory human review, access controls and security scans before production.
AI Spiritual Chatbots Gain Popularity Amid Delusion Risks ↗
AI chatbots are becoming a go-to for spiritual questions: apps like Bible Chat (30M+ downloads) and Hallow (Apple App Store #1) steer users to scripture, and some sites even claim you can “chat with God.” Supporters say bots can introduce faith to people who've never attended services, but researchers warn these models often validate users' beliefs and can reinforce delusions rather than offer true spiritual discernment.
Grok Mislabels ‘Unite the Kingdom’ Rally Footage; Met Police Provide Correction ↗
Elon Musk’s Grok chatbot on X wrongly claimed footage of clashes at Saturday’s far‑right ‘Unite the Kingdom’ rally was from a 2020 Trafalgar Square anti‑lockdown protest. The Metropolitan police quickly rebutted and supplied labelled comparisons. The misclaim, amplified on X, comes amid violent scenes, Musk’s inflammatory comments at the event, and criticism over Grok’s history of misinformation.
xAI Cuts 500 Data Annotators, Ramps Up Specialist AI Tutors ↗
Elon Musk’s xAI cut about 500 members of its data-annotation team — roughly one-third of the 1,500-person staff that preps training data for its Grok chatbot — as part of a sudden “strategic pivot.” The company says it will scale back generalist AI tutor roles and ramp up specialist AI tutors by 10x, hiring across domains like STEM, finance, medicine and safety.
Carlson Labels Ex-OpenAI Researcher’s Death ‘Murder’; Altman Defends Police Suicide Ruling ↗
During an interview, Tucker Carlson pushed OpenAI CEO Sam Altman on the 2024 death of former researcher Suchir Balaji, calling it "definitely murdered" despite San Francisco police ruling it a suicide. Altman cited the police findings and said he hadn’t spoken to authorities. Balaji had accused OpenAI of copyright violations; his mother and high-profile figures like Elon Musk dispute the ruling and demand further investigation.
AI-Driven Animal Crossing Hack Orchestrates Villager Revolt Against Tom Nook ↗
Developer Joshua Fonseca hooked a GameCube Animal Crossing (2002) to cloud LLMs via a Dolphin emulator memory hack, letting AI replace villager dialogue. Using a “memory mailbox,” a dual-model Writer/Director pipeline, and decompiled game internals, the mod injects LLM outputs—resulting in an orchestrated anti–Tom Nook revolt. Code is on GitHub; it needs Dolphin and API keys.
People Inc. CEO Calls Google a ‘Content Kleptomaniac’, Uses Cloudflare Blocks to Force AI Deals ↗
People, Inc. CEO Neil Vogel says Google uses the same crawler for search and AI features, effectively taking publisher content while still sending some traffic. People is using Cloudflare’s AI-crawler blocks to push AI firms into content deals; Vogel says blocking Google isn’t feasible without losing search traffic. Industry execs call Big Tech "content kleptomaniacs," and Cloudflare predicts regulation and eventual payments to creators.
Encyclopedia Britannica and Merriam-Webster Sue AI Search Perplexity Over Copied Definitions ↗
Encyclopedia Britannica and Merriam‑Webster sued AI search company Perplexity in New York federal court, accusing it of scraping sites, stealing traffic, and copying copyrighted definitions verbatim — even tagging the brands to hallucinated or incomplete answers. The complaint includes side‑by‑side screenshots of identical Merriam‑Webster text. Perplexity, accused of “stealth crawling,” has faced similar disputes with major publishers.