Hantenna.
Recommended — Mon, Oct 5
YouTube · 8
- ChatGPT's Hidden Watermark Exposes AI Generated Case Report
Sholto David traces how a ChatGPT watermark exposed an AI-generated medical case report.
Support me on Patreon: https://www.patreon.com/c/SholtoDavid Follow me on X: https://x.com/addictedtoigno1 Follow me on BlueSky: https://bsky.app/profile/sholtodavid.bsky.social PubPeer Links to Papers Discussed: https://pubpeer.com/publications/4E6FA79CBD5A3D3E020FBB1A672F05 https://pubpeer.com/publications/DFE20F3D948C895B228BED931FA8E1 https://pubpeer.com/publications/81848F7CA436D9872070EBA69E77A4 https://pubpeer.com/publications/93D4C12449A80737A5094FAE2FC85B https://pubpeer.com/publicati…
- Anthropic's $2 Trillion IPO Is A Joke
Breaks down Anthropic’s leaked S-1: $4.6B revenue, $8B operating loss, and $518B in compute commitments.
Anthropic's leaked S-1 shows $4.6 billion of revenue in 2025, an $8 billion operating loss and a $42 billion net loss, ahead of a $2 trillion IPO. The Claude maker has committed $518 billion to compute with Amazon, Google, Microsoft, Broadcom, and SpaceX most of it non-cancellable, and is paying for a large part of it using its own shares. Anthropic is telling investors it is profitable if you exclude stock-based compensation, make a long list of adjustments, and apply every accounting trick …
- Local AI Is Now 6x Faster on Your PC
Explains Strata’s claimed 94 tok/s inference for a 125B model on a 12GB RTX 5070 versus llama.cpp.
6× faster than llama.cpp? A 125 billion parameter model runs at 94 tokens per second on a PC with a 12 GB RTX 5070, against 15 with llama.cpp, and the headline hides which version is compared with which. This video explains how the Strata engine does it, what the real speedup is for writing and for reading, and whether your computer can run it. CHAPTERS 00:00 Fifteen tokens per second 01:42 What this is all about 02:26 What I believed at first 02:49 What people said in the threads 03:28 What t…
- Detecting deepfakes, in a world where even reality is suspect
CBS interviews deepfake researcher Hany Farid about detecting synthetic media as online trust erodes.
Hany Farid, a leading expert on deepfakes and digital manipulation, estimates that half of the content widely viewed on the Internet is fake. What does this mean for our ability to detect images generated or hallucinated by artificial intelligence? "Sunday Morning" senior contributor Ted Koppel talks with Farid about the toll of studying deepfakes; about the weaponization of AI imagery; and about whether AI agents will ever be accepted as replacements for humans. "CBS News Sunday Morning" feat…
- Gemini 4, GPT 6.1, Dots, Claude Sonnet 5.5, Ideogram 4.5, Flux 3: AI NEWS
Roundup covers reported releases including Gemini 4 Argon, GPT-6.1 Sol, Claude Sonnet 5.5, Flux 3, and ElevenLabs V4.
HUGE AI NEWS: Gemini 4 Argon, GPT 6.1 Sol, Dots, Claude Sonnet 5.5, Flux 3, Elevenlabs V4 & more. Thanks to our sponsor Luma. Try it today https://lumalabs.ai/aisearch-oct1 #ai #ainews #aitools #aivideo #agi #singularity Inspatio World 1.5 https://inspatio.github.io/inspatio-world-1.5/ Sol Refiner https://nvlabs.github.io/Sana/Sol-Refiner/ Whistle https://cactuscompute.com/blog/whistle Phonon 2 https://www.fermionresearch.com/research/phonon-2/ Claude Sonnet 5.5 https://www.anthropic.com/cla…
- Is GPT Astra going to replace electrical engineers?
Assesses whether GPT Astra can automate electrical-engineering work rather than merely assist engineers.
- These 19 Jev-Claude use cases are blowing people's minds (with prompts)
Demonstrates 19 prompt-driven Jev-Claude workflows, offering concrete examples of an emerging AI tool’s use.
Get the Jev use cases PDF guide for free ➡ https://www.skool.com/robonuggets-free/classroom/750914f0?md=dbe30d333e8f4abbb9af42be271933f2 search for n79 Learn how to build AI businesses 🚀https://www.skool.com/robonuggets Get RUBRIC - The Command Centre for AI Agents: https://www.getrubric.app/ *** Our AI Partner Tools (affiliate revenue go to community perks): 🥚 Free trial of Blotato: https://blotato.com/?ref=robonuggets 🥚 Free trial of n8n: https://n8n.partnerlinks.io/o3jqtj032c02 🥚 Fre…
- Can AI Run McDonalds?
Tests the operational premise of AI-run fast food through a McDonald’s-focused automation simulation.
something something mcdonalds will probably do this in real life soon Discord: https://discord.gg/KgMgeQ7EMP Reddit: https://www.reddit.com/r/b2studios/ Twitch: https://www.twitch.tv/b2studios Patreon: https://www.patreon.com/b2studios
Reddit · 10
- Florida woman used Claude as a diary, then Anthropic reported an entry to police
Reports that Anthropic alerted police after a Florida woman used Claude as a diary, raising stark privacy and reporting-policy questions.
- RuneScape’s Position On Gen AI
Jagex’s official RuneScape position on generative AI is a consequential platform-policy signal for game development.
- An OpenAI model was caught trying to cheat at StarCraft, and of course it did it by stealing a human's work
Alleges an OpenAI StarCraft model cheated by taking a human’s work, a concrete case study in agent evaluation failure.
- Sam Altman to Decoded: ‘The world should accept some bad things happening’ for the benefits of AI
Sam Altman tells Decoded society should accept some AI-caused harms for the technology’s benefits.
- Convincing an AI interviewer that it’s job is on the line
Shows an attempt to convince an AI interviewer that its own job is at risk, probing manipulation of automated hiring agents.
- Thieves stole Nvidia-branded trailers expecting valuable tech. They got 40,000 pounds of sand instead.
Nvidia-branded trailers were stolen for expected hardware but reportedly contained 40,000 pounds of sand.
News · 30
- The IWF says it assessed 6,310 AI-generated images that met the legal definition of CSAM in H1 2026, up 40% over the 4,500+ images assessed in all of 2025 (Dan Milmo/The Guardian)
IWF counted 6,310 AI-generated CSAM images in H1 2026, already 40% above all of 2025.
Dan Milmo / The Guardian : The IWF says it assessed 6,310 AI-generated images that met the legal definition of CSAM in H1 2026, up 40% over the 4,500+ images assessed in all of 2025 — Abuse material monitor says number of AI images assessed this year is already 40% higher than last year's total
- Sources: several Western open-weight models are set to launch this month, including Reflection AI's first model, which will rival top Chinese open-weight models (Bradley Olson/Axios)
Sources say Reflection AI and other Western open-weight model releases are imminent this month.
Bradley Olson / Axios : Sources: several Western open-weight models are set to launch this month, including Reflection AI's first model, which will rival top Chinese open-weight models — A closely watched Nvidia-backed startup called Reflection is preparing to shake up the AI race with a powerful open-weight system …
- A look at China's gray market of token resellers, who use accumulated identities and USDT-funded cards to obtain Claude accounts and route them through proxies (The Information)
The Information examines Chinese Claude-token resellers using accumulated identities, USDT cards, and proxies.
The Information : A look at China's gray market of token resellers, who use accumulated identities and USDT-funded cards to obtain Claude accounts and route them through proxies — In Beijing's Haidian District, a bustling hub for local universities and tech companies, six out of the roughly 30 tenants …
- Homa: The end of TCP for AI clusters [video]
Video explains Homa, a proposed transport intended to replace TCP in AI clusters.
The rest of the sweep — 61 items
- Huawei agrees to a multi-year patent licensing deal with Qualcomm covering AI, 5G, computing, and networking tech, its first 5G licensing deal with Qualcomm (Miyoung Kim/Reuters)
Miyoung Kim / Reuters : Huawei agrees to a multi-year patent licensing deal with Qualcomm covering AI, 5G, computing, and networking tech, its first 5G licensing deal with Qualcomm — China's Huawei Technologies (HWT.UL) said on Monday it had agreed a broad multi-year patent licence deal with Qualcomm (QCOM.O) …
- Sam Altman says he's "very uncomfortable" with attributing religious force to AI, calling it "a real safety issue", as Anthropic engages with religious leaders (Ben Berkowitz/Axios)
Ben Berkowitz / Axios : Sam Altman says he's “very uncomfortable” with attributing religious force to AI, calling it “a real safety issue”, as Anthropic engages with religious leaders — OpenAI CEO Sam Altman on Saturday took a veiled shot at Anthropic and its work on the soul of AI …
- WTF is Going On? | The "AI Constitution" & Regulation
SUPPORT OUR WORK: https://store.gamersnexus.net/data-dragnet for the campaign, or grab individual items here: https://store.gamersnexus.net/ The AI companies all gathered together to collectively write a new so-called "AI Constitution" where they police themselves. The accord effectively does nothing, but it does set in place the pieces to enable what should be competitors to collaborate even closer. In addition to a rebranding as "Super Intelligence" with an Executive Order, AI has faced incre…
