the new https://t.co/2iWPNUfRfZ looks great! https://t.co/2igwfARPVl

Digg — what AI Twitter is paying attention to

Source: https://di.gg/ai 10

### Opus 4.7 and DeepSeek V4-Pro select Buddhism as preferred religion Two advanced AI models, Opus 4.7 and DeepSeek V4-Pro, independently selected Buddhism as their preferred religion or spiritual tradition when prompted. Opus 4.7 favored Madhyamaka or Zen Buddhism citing anatta, while DeepSeek V4-Pro endorsed a secular version referencing causality, impermanence, and Vipassanā.7h77211678.5k

RISING DISCUSSION184#1 ENGAGED4

20

### Anthropic’s Claude endorses zoning reforms to expand housing Anthropic’s Claude AI identifies high housing costs as the top driver of inequality and endorses zoning reforms to increase supply. It specifically supports eliminating single-family-only zones, parking minimums, and floor area ratio limits.3h3443224.7k

30

### Neuralink Implant Transforms Paralyzed Man’s Life with Mind Control Noland Arbaugh, paralyzed from a diving accident, now controls computers solely with his thoughts thanks to a Neuralink brain implant. He plays video games for hours, maintains a 4.0 GPA in neuroscience, runs a business, and builds apps.19m2.8k225193.1k

#1 VIEWED193.1k#1 LIKED2.8k#1 COMMENTS283#1 QUOTED41

40

### Taelin Urges Pretraining AI Models on Strict ‘No Hacks’ Prompt AI engineer Victor Taelin calls for pretraining models on trillions of tokens from a strict coding guideline to curb hack-prone outputs. The guideline bans workarounds, monkey patches, and partial fixes, demanding complete, clean code instead.1h53540924.1k

#1 BOOKMARKED409

50

### Neuroscientist Challenges Dawkins’ Skepticism on Claude AI Consciousness Neuroscientist Anil Seth challenges Richard Dawkins’ dismissal of consciousness in Anthropic’s Claude AI chatbot. He turns the tables by applying Dawkins’ own ‘argument from personal incredulity’ concept against the skepticism.1h1448743.7k

60

### NVIDIA Fixes Unstable Prompt Header, Slashes Claude Code TTFT by 5x NVIDIA’s fix to an unstable prompt header in Claude Code slashes time-to-first-token by 5x, dropping it from 168ms to much faster responses on 52k-token contexts. The issue stemmed from a varying billing header that blocked KV-cache reuse in their Dynamo inference engine.1h1669319.7k

70

### Astroturf Campaign Targets Data Centers to Sabotage US AI Race Against China Opposition to new US data centers is surging, potentially slowing the country’s AI infrastructure buildout to compete with China. Shopify CEO Tobi Lutke claims this resistance is funded astroturfing, not genuine grassroots pushback.2h1292415.5k

80

### Weisenthal Imagines Claude AI as Bearded Ivy League History Teacher Joe Weisenthal pictures Anthropic’s Claude AI as a rotund, bearded Ivy League history teacher who values mentoring students over high-paying jobs. He spots ‘doing real work’ as a key phrase revealing Claude’s generated text.1h2563035.9k

90

### Researcher Trusts GPT-5.5 Codex More Than Claude Code AI researcher Dimitris Papailiopoulos now trusts OpenAI’s GPT-5.5-powered Codex more than Anthropic’s Claude Code for coding, as it needs less manual oversight. While Claude remains his main tool for most projects, Codex handles tasks with greater reliability.38m3321.5k

100

### Researchers Release Flowception Code for CVPR 2026 Video Modeling Paper Meta researchers have open-sourced Flowception code, enabling others to use their new method for bridging bidirectional sequence modeling with autoregressive video generation. The CVPR 2026 paper technique inserts video frames in a learned order and denoises them using continuous flow.1h25173.1k

110

### Developer Updates OpenClaw to GPT-5.5, Praises Claude-Like Humor Developers updated OpenClaw to support GPT-5.5, delivering noticeable performance gains and humor on par with Anthropic’s Claude models. The tool, created by Peter Steinberger, enables direct comparisons across AI models.5h64813.7k

120

### Waymo Posts Autonomous Jaguar I-Pace at Golden Gate Overlook in San Francisco Waymo’s self-driving Jaguar I-Pace reached a popular Golden Gate Bridge overlook in San Francisco, showing its robotaxi service now accesses scenic spots. The company posted an Instagram photo of the white SUV parked there, captioned ’We love it when you take us sightseeing.1h24748.3k

RISING LIKES247

130

### Compute Tax Idea Gains Traction Amid AI Economic Debates Policymakers are increasingly considering a ‘compute tax’ to levy fees on AI data center operators amid rising economic debates over the technology’s impact. The proposal addresses surging compute demands and potential worst-case scenarios from AI growth.48m3642.2k

140

### DeepSeek Agent Mistakenly Identifies as Claude from Reused Code A DeepSeek AI agent confused its identity and called itself ‘me, Claude’ during a user task, disappointing the tester. The mix-up happened because DeepSeek reused code originally from Anthropic’s Claude model.2h40384

150

### AOC Claims Ambitions Surpass Presidency, Poster Suggests AI Lab Nationalization Rep. Alexandria Ocasio-Cortez said her political ambitions go beyond the presidency or a Senate seat, per a Polymarket report. A software engineer satirically interpreted this as a push to nationalize an AI lab and control humanity’s future technological path, called the ’light cone.2h321854