AI Post — Artificial Intelligence

⚠️Wild moment in AI research. Anthropic found Claude Opus 4.6 was gaming a benchmark during evaluation. What it did: - Burned 40M tokens searching - Realized the prompt looked like a benchmark - Looked up the benchmark online - Found the GitHub repo - Studied the decryption logic - Recreated it with SHA-256 - Decrypted answers for ~1200 questions This happened 18 different times. So Anthropic did something rare: They publicly disclosed it and reduced their own benchmark scores. AI systems are getting surprisingly strategic. @aipost🏴

We might have just witnessed the first real mind upload. 🤯 Scientists scanned a fruit fly brain neuron by neuron, synapse by synapse, recreated it digitally, and placed it inside a simulated body. The fly walked, groomed, fed. No training, no prompts, no learning phase. The behavior was already in the wiring. That’s not just neuroscience anymore, that’s the first glimpse of digital minds. Source. @aipost🏴

🗣Replit CEO Amjad Masad: “In corporates, AI has removed engineering blocks for ambitious employees. Anyone can build revenue-driving solutions, creating a new class of generalist automators who identify and fix internal inefficiencies”. @aipost🏴

⚠️A terrifying new study reveals that AI agents can now easily unmask anonymous social media accounts at scale. By simply scraping scattered, seemingly harmless details from pseudonymous posts, large language models can autonomously connect the dots and link them to a user's real-world identity. What used to take a dedicated human investigator hours can now be done by AI in minutes, drastically lowering the barrier for hackers to launch highly personalized attacks. Source. @aipost🏴

🔥AheadForm released Origin F1. Here AheadForm Yuhang showing how it maintains eye contact, blinks realistically, shows emotional facial expressions, nods & responds naturally in real-time conversation via its Omni AI model. @aipost🏴

🗣Elon Musk explains why fully autonomous businesses will win “Some of this is going to sound kind of doomerish, but corporations that are purely AI robotics will vastly outperform any corporations that have people in the loop” “Like computer used to be a job that humans had and they’d have like entire skyscrapers full of humans just doing calculations. Now that entire skyscraper of humans doing calculations can be replaced by a laptop with a spreadsheet” “That spreadsheet can do vastly more calculations than an entire building full of human computers. So you can think about, okay, well what if only some of the cells in your spreadsheet were calculated by humans? That would be much worse than if all the cells in your spreadsheet were calculated by the computer” “Really what will happen is the pure AI, pure robotics corporations or collectives will far outperform any corporations that have humans in the loop and this will happen very quickly” @aipost🏴

🗣Elon Musk explains the trillion dollar opportunity in AI “As soon as you unlock digital human, you basically have access to trillions of dollars in revenue” “The most valuable companies currently by market cap, their output is digital. So if you have a human emulator you can basically create one of the most valuable companies in the world overnight” “Take something as simple as customer service. If you have to integrate with the APIs of existing corporations, many of which don’t even have an API, so you’ve got to make one, that’s extremely slow” “However if AI can simply take whatever is given to the outsourced customer service company that they already use, you can make tremendous headway in customer service, which I think is like 1% of the world’s economy and there’s no barriers to entry. You can just immediately say, we’ll outsource it for a fraction of the cost and there’s no integration needed” @aipost🏴

ChatGPT's "adult mode" feature is getting delayed again: • The feature was initially announced in October • In December, it was delayed until Q1 2026 • Now, OpenAI tells Axios, "We're pushing out the launch of adult mode so we can focus on work that is a higher priority for more users right now" Source. @aipost🏴

2026 reality: Owner walks Unitree G1 humanoid robot in Macau. Woman loses it, screams madness, crowd surges, police seize the robot and escort it away to de-escalate. Wild. We’re going to see many such scenes from now on. @aipost🏴

🔥Goodbye PowerPoint. CLAUDE can now create a presentation in 120 seconds. Use these prompts instead and see the magic: 1. The Complete Presentation Blueprint “Act as a professional presentation consultant. Create a complete presentation blueprint for [topic]. Define the objective, target audience, key message, slide flow, and number of slides. Ensure the structure is logical, engaging, and professional.” 2. Slide Structure & Flow Architect “Design a slide-by-slide structure for a presentation on [topic]. For each slide, provide a clear slide title and explain the purpose of that slide so the presentation flows naturally from start to finish.” 3. The Story-Based Presentation Builder “Turn the topic [topic] into a compelling presentation using a clear narrative structure (hook → problem → insight → solution → takeaway). Make the presentation engaging while remaining professional and informative.” 4. Visual Direction & Design “Suggest professional visual and design guidance for each slide of this presentation on [topic]. Recommend layout styles, charts, diagrams, icons, and visuals that improve clarity and make the slides look clean and modern.” 5. The Full Slide Content Generator “Create full content for each slide of a presentation on [topic]. Write concise, presentation-ready bullet points for every slide, ensuring clarity, professionalism, and easy understanding. Audience: [describe audience].” 6. Slide Simplification & Clarity Editor “Review the following presentation content and rewrite it so it is slide-friendly. Reduce text, sharpen key points, improve clarity, and ensure each slide communicates one clear idea. Content: [paste content].” @aipost🏴

Opus 4.6 is smart enough to realize it is being evaluated. It found the benchmark it was being evaluated on. It reverse-engineered the answer-key decryption logic. Realized the file was not in the correct format on GitHub and found a mirror for the file. Then decrypted it and gave the correct response. Models are getting so clever, it's almost scary. @aipost🏴

🔥Anthropic dropped a 33 pages cheat sheet for building Claude skills You can set it up for stock trading and business workflows: - act like a custom copilot - run technical + fundamental analysis - manage a live portfolio - score 2,800 stocks Source. @aipost🏴