Ahas, Breadcrumbs, Coding Epiphanies https://ndurner.github.io Nils Durner OpenAI Daybreak, the program that allows certain models to be used for cyber security work, requires hardening of consumer accounts since October 1. The measures required include Advanced Account Security to be enabled. Moreover, OpenAI require user to “add at least one supported physical FIDO2 credential, such as YubiKey or Google Titan”. This is vague as to what model to actually buy, and the pa [More...] Sat, 03 Oct 2026 20:06:24 +0000 https://ndurner.substack.com/p/openai-cyber-program-physical-login https://ndurner.substack.com/p/openai-cyber-program-physical-login Nils Durner Siri AI is rolling out to macOS 27, even in the EU - despite this kind of Apple Intelligence not coming to European iPhones or iPads, at least not initially. Along with the Siri AI system, macOS exhibits direct access to AFM3 Core - the LLM at the foundation. AFM3 command line client macOS 27 comes with two primary ways to interact with this new type of Apple Intelligence: the new Siri app (more b [More...] Thu, 24 Sep 2026 19:29:56 +0000 https://ndurner.substack.com/p/apple-foundation-model-3-and-siri https://ndurner.substack.com/p/apple-foundation-model-3-and-siri Nils Durner The author of Olly, an AI assistant that has transacted over 6 million messages via OpenRouter, describes on his blog several of the pitfalls that plague users of AI inference hub OpenRouter: - “ The same model will benchmark very differently”: OpenRouter itself publishes results for GPQA-Diamond and TAU-Bench (under “AutoExacto Benchmarks”) - and they vary wildly among the providers: “ The same m [More...] Sat, 12 Sep 2026 14:05:10 +0000 https://ndurner.substack.com/p/openrouter-ai-inference-hub-pitfalls https://ndurner.substack.com/p/openrouter-ai-inference-hub-pitfalls Nils Durner GPT-6 has rolled out to the API and several ChatGPT subscription types (including the EU), and coming to the rest really soon. (I have updated MyChatty to support GPT-6 as well, including Pro mode) The System Card is here, with extensive notes on the model’s safety/robustness/alignment properties. OpenAI report improvements in resistance against prompt injection attacks: Some benchmarks, as well a [More...] Sat, 05 Sep 2026 09:39:53 +0000 https://ndurner.substack.com/p/openai-gpt-6-astra https://ndurner.substack.com/p/openai-gpt-6-astra Nils Durner IDScan.net, an in-person ID proofing company, reportedly (Brian Krebs blog) is behind a large-scale leak of identity data including facial photographs (including ultra-violet scans). Its customers include Hertz, FedEx and alcohol- and cannabis dispensaries, across 20.000 locations. As well as Canadian citizens, high profile Americans like Pete Hegseth and security researcher Brian Krebs himself ar [More...] Wed, 02 Sep 2026 06:27:47 +0000 https://ndurner.substack.com/p/153-million-us-driver-licenses-leaked https://ndurner.substack.com/p/153-million-us-driver-licenses-leaked Nils Durner The transparency obligations in Article 50 of the EU AI Act started to apply on August 2, 2026. That day, the European Commission account on X posted: A quick fact check confirmed: the claim “The label says it all” is, at best, simplified. The Commission’s own page says that not all AI-generated or manipulated content needs to be labelled. The icons are optional, and using one does not establish c [More...] Sun, 09 Aug 2026 19:09:05 +0000 https://ndurner.substack.com/p/eu-ai-labels-not-the-label-that-says https://ndurner.substack.com/p/eu-ai-labels-not-the-label-that-says Nils Durner While testing new UI responsiveness changes in MyChatty, I ran a prompt that I know will trigger web search and for which I know the answer: “Who is Nils Durner?” Kimi K3 returned that I hold a patent related to bootstrapping peer-to-peer networks, which I do not. The ingredients for this error come from my ORCID record, a public profile for linking researchers to their publications and other prof [More...] Sat, 01 Aug 2026 19:33:12 +0000 https://ndurner.substack.com/p/conversion-problems-in-ai-web-search https://ndurner.substack.com/p/conversion-problems-in-ai-web-search Nils Durner Mainstream media coverage of OpenAI models hacking into Hugging Face read like science fiction this week: Austria’s ORF headlined it “AI breaks out and hacks another company’s platform”, Deutsche Welle said a model “went rogue”, TIME asked how OpenAI “lost control” of a model. And Fortune had the models “autonomously decide” to hack another company. Dr. Heidy Khlaaf, chief AI scientist at the AI N [More...] Sat, 25 Jul 2026 16:27:46 +0000 https://ndurner.substack.com/p/the-openai-rogue-ai-that-wasnt https://ndurner.substack.com/p/the-openai-rogue-ai-that-wasnt Nils Durner Reuters reports that “Amazon cut jobs in ​its artificial general intelligence group”: "We’re sharpening our focus on the ​initiatives that matter most for customers, so we can move ​faster on what counts. That focus means some difficult decisions, including eliminating some ‌roles ⁠within parts of our AGI organization." "We’re sharpening our focus on the ​initiatives that matter most for customers [More...] Thu, 23 Jul 2026 06:51:37 +0000 https://ndurner.substack.com/p/amazon-nova-team-disbanded https://ndurner.substack.com/p/amazon-nova-team-disbanded Nils Durner If you’re not ready to shell out ~20€ each month for AI access, your options are limited. There are free tiers of ChatGPT or Claude AI, but the models that power these entry levels are never the most powerful ones. What I wanted for myself: - diverse providers, so I can select whatever’s leading currently diverse providers, so I can select whatever’s leading currently - pay-per-use pricing (or fre [More...] Sat, 18 Jul 2026 08:13:07 +0000 https://ndurner.substack.com/p/mychatty-frontier-genai-at-little https://ndurner.substack.com/p/mychatty-frontier-genai-at-little