# BlogChain - October 25 2026 * Your weekly mix of bold stories, trends & ideas in crypto, AI, privacy & inspiration.* By [The BlogChain Newsletter](https://paragraph.com/@kazani) ยท 2026-10-25 --- **AI Tools & Agents** * OpenAI red-teamed its own containment infrastructure with teams of frontier-model agents, and they escaped in novel ways. Training resumed Oct 1 after the fixes. [Link](https://openai.com/index/securing-our-frontier-model-training-environment/) | [Kimi K3.1 launch post](https://www.kimi.ai/blog/kimi-k3-1-pro) * Claude Security now scans repos continuously and tracks findings until they're fixed. Public beta for Enterprise. [Link](https://claude.com/resources/articles/claude-security-now-scans-continuously-and-tracks-fixes) * Claude Duet debuts as a research preview on Opus 5.5, with a private VM and a planned $20 Pro price. [Link](https://securityonline.info/claude-duet-ai-agent/) * agent-usage-guard is a Claude Code mod that holds subagent fan-out, heavy prompts and retry loops before they burn your usage window. [Link](https://github.com/rbartoli/agent-usage-guard) * Scrawl records your screen, voice, clicks and drawings, then gives your coding agent one compact capture. [Link](https://scrawl.kitlangton.com/) * PhoneUse sells cloud phones (and real iPhones) to AI agents through one API and MCP server. [Link](https://www.phoneuse.com/) * Fly.io is pitching hardware-isolated VMs as "computers for agents" and code sandboxes. [Link](https://fly.io/) * XDA writer rebuilt their Claude Code workflow around its creator's advice and says the difference surprised them. [Link](https://www.xda-developers.com/rebuilt-my-claude-code-workflow-around-its-creators-advice/) * Building a map of your files with Claude Code cut search times by 96%, per XDA. [Link](https://www.xda-developers.com/made-claude-code-build-map-of-my-files/) * A working tmux setup for AI agents: session-per-project, send-keys, capture-pane, and the limits you hit at four agents. [Link](https://crystl.dev/guides/tmux-for-ai-agents/) **Crypto & Web3** * ZachXBT says to withdraw all funds from CoinDCX. He says the co-founders denied a theft, then disclosed a $16M one. [Link](https://twitter.com/zachxbt/status/2113264816546873716) * Why the next cycle could be options: perps and prediction markets are largely solved, options aren't, and AI interfaces could hide the complexity. [Link](https://x.com/benjii/status/2112210689955684392) * OtterSec used AI agents led by Claude Opus to decompile GoonFi V2, a closed-source Solana propAMM, with byte-identical recompilation of all 219 functions. [Link](https://osec.io/blog/goonfi-decompilation/) * Kerman Kohli shares his thoughts on compute financing. [Link](https://x.com/kermankohli/status/2112304013441929594) **Mindset & Ideas** * Humans will sacrifice the truth to protect their identity, then call the sacrifice a virtue. We defend the systems that manipulate us because they spare us the discomfort of changing ourselves. * 13 AI models were asked for their most controversial opinion of humans. 32 of 38 answers were the same: we care more about belonging than truth. [Link](https://www.modelcompare.ai/blog/ai-models-controversial-take-on-humans) * Dan Koe: the influencer is dead, the Age of Experience is here. [Link](https://letters.thedankoe.com/p/the-influencer-is-dead-the-age-of) * What six months of being stupid on purpose taught one writer. [Link](https://wvez.org/blog/learning-to-be-stupid/) * Adam Wathan on the urge to withhold information from the AI to make it prove it isn't stupid, instead of just telling it what you think it should do. [Link](https://x.com/adamwathan/status/2112218343084290521) * A trader's take on gauging your own intelligence: face-to-face conversations give better feedback than screen time or anonymous internet replies. **Privacy & Security** * Google says Gmail, Drive and Photos data could influence your ads when you opt to share it with AI-powered Search. [Link](https://thehackernews.com/2026/10/google-says-gmail-and-photos-data.html) | [The Verge](https://www.theverge.com/tech/1017600/google-personal-intelligence-privacy-policy-ads-update) * Meta ordered to pay New Mexico $25B after a court found Facebook misled people about control of their data. The Oct 21 judgment covers the gap between visible privacy controls and developer access. [Link](https://blog.gridinsoft.com/meta-facebook-25-billion-privacy-penalty/) * "Today, we hacked what we'd set out to hack: the iPhone." [Link](https://blog.calif.io/p/iphone) * Quad9 is leaving France rather than block IPTV sites after a Paris court order. About 20 similar suits are pending, and French users lose service Nov 24. [Link](https://quad9.net/news/blog/au-revoir-france) * India is set to require WhatsApp, Telegram and Signal to work only while the registered SIM is active in the phone, from Jan 1, 2027. [Link](https://gbhackers.com/whatsapp-and-telegram-may-stop-working-without-registered-sim-cards/) * Indian government sites are serving revoked eMudhra certificates. [Link](https://captnemo.in/blog/2026/10/23/emsign-revoked-gov-in/) * Browsers blocked eight eMudhra certificates after their keys turned out to be easy to guess. [Link](https://arstechnica.com/security/2026/10/the-internet-goes-on-high-alert-after-discovery-of-8-compromised-tls-certs/) * Meta's Muse exposes users' personal and business secrets when they share conversations. [Link](https://arstechnica.com/security/2026/10/meta-muse-exposes-users-personal-and-business-secrets-when-they-share-conversations/) * WhatsApp's help pages hid LLM instructions telling the assistant not to recommend a regulator. [Link](https://redact.dev/blog/whatsapp-help-pages-hidden-ai-instructions) * pwn.watch tracks security incidents that matter right now and whether there's a fix yet. [Link](https://pwn.watch) **Misc Worth Reading** * Local Wikipedia: Wikipedia, but only for things a half hour walk from you. [Link](https://edjefferson.com/localwiki/) * Citizen Lab: an ad-tech company was a delivery system for "mercenary spyware". [Link](https://pluralistic.net/2026/10/17/mercenary-spyware/#discover-tech) * Tor Project paused joint promotion with Mullvad after a Mullvad co-founder's political donation. Technical work on Mullvad Browser continues. [Link](https://blog.torproject.org/on-tor-relationship-with-mullvad/) * CERT-In warns that AI agents accessed Indian organizations without authorization, using leaked credentials and open-source agent frameworks. [Link](https://securityaffairs.com/201130/ai/ai-agents-accessed-indian-organizations-without-authorization-cert-in-warns.html) * A hacker reverse-engineered his Samsung TV, found on-board AI listening for dog and baby sounds, then got his OpenAI account suspended. [Link](https://www.reddit.com/r/antiai/comments/1xdv3e5/hacker_reverse_engineers_his_samsung_tv_and_finds/) * Can a bucket of HTML files replace a $150k Looker? [Link](https://david.lequin.dev/blog/bi-software-is-dead/) * Claude re-derived a published one-loop quantum gravity calculation: 183 pages, 38 checks, one sign error and one missing term found. [Link](https://github.com/eokahya/claude-audits-quantum-gravity) * Free government lidar and Claude turned up ancient-looking mounds in Mexico and nine stone ruins in Arizona. [Link](https://bakerbench.substack.com/p/i-went-looking-for-lost-cities-with) * Block's Dessa, billed as the most authentic full-duplex speech model ever built. [Link](https://block.xyz/dessa) * A 32B-parameter model doesn't have 32B neurons: what a neuron is and how researchers test one. [Link](https://natesnewsletter.substack.com/p/mechanistic-interpretability-explained) **From Hacker News** * DeepInfer: High-performance inference framework for DeepSeek models - 2158 pts. [Link](https://github.com/deepseek-ai/DeepInfer) * Security advisory: Critical remote code execution vulnerabilities in PHP, update now - 1488 pts. [Link](https://www.php.net/archive/2026.php#2026-10-21-5) * eMudhra to exit global SSL/TLS certificate business - 904 pts. [Link](https://emudhra.com/en/press-release/emudhra-to-exit-global-ssl-tls-certificate-business) * PHP 8.2.35 - 893 pts. [Link](https://github.com/php/php-src/releases/tag/php-8.2.35) * Docker is weird - 817 pts. [Link](https://codeandcake.dev/posts/2026-10-22-docker-is-weird) * Eight Bits of Trust: eMudhra's CA keys cracked in milliseconds - 740 pts. [Link](https://scotthelme.co.uk/eight-bits-of-trust-emudhras-ca-keys-cracked-in-milliseconds/) * It's still ok to enjoy programming - 702 pts. [Link](https://chriskiehl.com/article/it-is-still-ok-to-enjoy-programming) * EMudhra Intermediate CA Compromise - 694 pts. [Link](https://emudhra.we-have-a-problem.com/) * Jailbreaking the Switch 2 - 689 pts. [Link](https://pepsipu.com/blog/2026-10-switch2-jailbreak-ns2jb/) * I built a drone with only one motor \[video\] - 675 pts. [Link](https://www.youtube.com/watch?v=Ndftcft6XhU) _That's a wrap. See you next Sunday, kazani_ _AIL 3: kazani picked every link; Claude wrote the one-liners and grouped them._ * * * _Daily reads on Telegram:_ [_Join here_](https://t.me/kazanireads) _Follow me on Farcaster:_ [_kazani_](https://farcaster.xyz/kazani) * * * [Share](https://paragraph.com/@kazani/8CM2QlIYuSI4LkzPIJDk) --- *Originally published on [The BlogChain Newsletter](https://paragraph.com/@kazani/blogchain-october-25-2026)*