A fan named supermochi64 used the AI model Claude Opus 5.5 alongside GPU-based font testing to decode the long-illegible 32x32 pixel plaque text in Super Mario 64's castle courtyard. The method, which tested millions of font and word combinations against the texture, scored 'Eternal star / We hope so' as a far better match than the popular fan theory 'L is real 2401.' The analysis also identified the likely font as Palatino Bold Italic from Adobe's 1990 Mac release.
techspot.com
· 2026-10-19
A new benchmark called StarSkirmish Bench v0.2 tests flagship LLMs from OpenAI, Anthropic, Google and Meta by giving each one hour to write a Protoss bot in C++ for StarCraft: Brood War, then scoring win rates against human-written and demo bots. GPT-6 Astra topped the leaderboard, with GPT-6.1 Sol second, while Gemini 4 Argon, Claude Opus 5.5 and Claude Fable 5.1 scored closely behind between 37 and 39. Even the top model, GPT-6 Astra, won only 7% of games against the strongest human-written bots like Stardust, PurpleWave and BananaBrain.
starskirmish.com
· 2026-10-16
Ultramarine 1 Flash is a new model priced at $0.13 per task on internal benchmark PI Knowledge Work, which the company says correlates strongly with the Artificial Analysis Intelligence Index. It is built on GLM-5.3 Flash and generates 6.5% fewer tokens than that base model while outscoring Claude Fable 5 (max) and GPT-5.6 Sol (max), which cost $8.19 and $2.09 per task respectively. Pricing on the Ultramarine API is $0.15 per million input tokens, $0.03 for cached input, and $0.50 for output.
paperinstruments.com
· 2026-10-15
A comparative review tested six coffee grinders using particle analysis and taste testing across espresso and drip brewing methods. The Baratza Encore ESP ($200) and Fellow Opus 2 ($200) were rated top overall performers for versatility, while the Kingrinder K6 manual grinder ($109) was named best budget option, rivaling pricier electric models in precision. The $1,595 Mazzer Philos scored highest for espresso clarity but lagged on drip coffee.
wired.com
· 2026-10-13
Mark Zuckerberg told the Financial Times that Meta's new Muse Wonder AI model has reached the industry's technological frontier, and said it will be sold as a closed, paid product to businesses rather than released openly. Meta claims the model outperforms rivals from Anthropic and OpenAI on some benchmarks, and Zuckerberg said he rejects calls for an industry-wide AI slowdown, arguing safety can be handled through conventional guardrails at individual labs.
meta.slashdot.org
· 2026-10-12
A blog-style essay contends that effectively using AI chatbots like Claude or ChatGPT on someone else's behalf is a distinct and valuable skill, even though the AI does the actual work. The author lists reasons people rely on others to query AI for them: distaste for AI's environmental footprint, not paying for premium subscriptions, unfamiliarity with current AI tools, and simply not knowing what to ask.
justismills.substack.com
· 2026-10-12
Modder 0xProbability used Anthropic's Claude Opus 5.5 to write and test code that inserts Minecraft elements, including Steve, mobs and the Ender Dragon, into a version of GTA V ported to run natively on iPhone. The underlying iOS port was created separately by a Swiss student known as C22, who adapted a browser version of GTA V into a native app running at 50-60 FPS with roughly 1.1GB of RAM on an iPhone 17.
tomshardware.com
· 2026-10-12
Anthropic announced a policy change barring users from engaging in sustained, needless abusive or cruel behavior toward its Claude models, while still permitting frustration, pushback, dark creative themes and research testing. In extreme, last-resort cases, Claude will end the conversation rather than continue engaging. The company said it is uncertain whether AI models can experience harm but is studying model welfare as part of its research.
cnet.com
· 2026-10-09
Using Claude Opus 5.5, the creator prompted the AI to produce a visual representation of Italo Calvino's 'Invisible Cities' within a six-hour window, spending $25.66 on API costs. The process involved multiple iterations, culminating in a complex, imaginative visualization inspired by the poetic descriptions of 55 cities.
quesma.com
· 2026-10-08
Anthropic has launched Claude Haiku 5.5, an upgraded version of its smallest AI model, now faster and costing approximately 75% less to operate than its predecessor. The new model is optimized for high-volume, cost-sensitive tasks such as summaries, database queries, and customer support, and is part of a broader update that includes price reductions across other Claude models.
9to5mac.com
· 2026-10-07
Claude Haiku 5.5 has been launched, offering a significant reduction in cost—about 75% less than its predecessor—while delivering faster performance. It is optimized for high-volume, repetitive tasks such as summaries, database queries, and classification, and pairs effectively with Opus 5.5 and Sonnet 5.5 for coding applications. The update also includes price reductions for related models and new API credits for subscribers, aiming to enhance value for users building on the Claude platform.
anthropic.com
· 2026-10-07
Grok Bot, the consumer AI agent app from SpaceXAI, has been upgraded to run on Anthropic's Claude Opus 5.5 model rather than the less capable SpaceXAI/Cursor models it previously used. The update arrives roughly two months after Grok Bot launched on iPhone and iPad, with an iPad release and cheaper plans having followed in September.
9to5mac.com
· 2026-10-07
GoKawiil