Tech

Personal News I got accepted into the Claude Code hackathon last week. 13,000 people applied, 500 got in. The hackathon just wrapped up today and I built Robo.app . Check it out, and watch the 3 min pitch video ! Here’s what caught my attention from Feb 10-16, 2026: 🚀 OpenAI Hires OpenClaw Founder for “Personal Agents” Peter Steinberger joining OpenAI to build next-gen personal agents - Sam Altman personally welcomed him. OpenClaw hit 100K GitHub stars in 2 days (fastest any repo has ever done that), now at 197K+. Both Meta and OpenAI made offers. The fact they’re building a dedicated personal agents team, not folding him into ChatGPT, signals a new product line. OpenClaw will transition to an independent open-source foundation that OpenAI will continue to support. ( @sama ) 🏋️ Three Major Model Launches in One Week Qwen 3.5-397B released - 397B total params, 17B active via hybrid linear attention + MoE. 87.8 MMLU-Pro, 83.6 LiveCodeBench v6, 76.4 SWE-bench Verified. Decoding 8.6x-19x faster than Qwen3-Max. Apache 2.0, 201 languages. Already on OpenRouter . ( @Ali_TongyiLab ) MiniMax M2.5: Sonnet-level coding at 1/20th Opus cost - 230B MoE, 10B active. 80.2% SWE-Bench Verified, $0.30/M input tokens. OpenHands confirmed it’s the first open model to beat Sonnet on their benchmark. Stock jumped 15.7%. If you’re running coding agents, test this as your workhorse model. ( VentureBeat ) ByteDance Seed 2.0 / Doubao 2.0 - Claims GPT-5.2 and Gemini 3 Pro level at ~1/10th cost. Doubao app has 155M weekly active users. China’s frontier labs aren’t 6 months behind anymore. ( @QuanquanGu ) 🌐 WebMCP: Every Website Becomes an API for Agents Google and Microsoft co-authored a W3C spec called WebMCP - Websites can now declare capabilities as structured tools via navigator.modelContext. Two modes: declarative (HTML forms self-describe) and imperative (JS exposes custom tools). 89% token efficiency improvement over screenshot-based agents. Available in Chrome 146 Canary behind a flag. ( @aakashgupta ) Pietro Schirano built a DoorDash agent demo - Adds items to cart and checks out without ever parsing UI. This is what “agents that talk to apps” looks like. ( @skirano ) Links: Chrome blog | W3C spec draft ...

I am increasingly getting firsthand stories of fraud hitting people of all ages, tech-savvy or not. The FBI’s IC3 2024 report logged 859K complaints with $16B in losses. That number is up 33% from 2023. Here is what is actually happening and what to do about it. Threat vectors (ranked by prevalence) Phishing and smishing Still the number one attack vector by volume. Fake texts from a bank, Robinhood, Coinbase or “USPS” with seemingly urgent links. The Verizon DBIR consistently finds phishing in over 30% of breaches. Do not click links or call numbers from unsolicited messages. Go directly to the website or app instead. Hardware security keys are the strongest defense since phishing sites cannot intercept them. ...

As a heavy daily AI user, one thing is painfully clear: AI hits a ceiling fast when it doesn’t have the right context. This is especially true as we move into autonomous AI agents that can solve problems before you even ask. I’m working on this through an autonomous OpenClaw agent that runs on a server in my basement (more on that soon), but even for AI beginners, I wanted to illustrate this with a practical example. ...

Here’s what caught my attention from the past 48 hours (Feb 6-7, 2026): 🤖 Opus 4.6: The Agent Model Arrives Claude Opus 4.6 launches with agent teams, fast mode, and 1M context - SOTA on agentic coding benchmarks. Can plan before acting, catch its own mistakes, and coordinate parallel agents. This is the model designed for tasks that span hours, not minutes. ( @claudeai ) Agent teams are experimental but available now - Set CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS and spawn multiple agents that coordinate via shared task lists and a mailbox system. I’ve been running these all weekend. ( @bcherny ) Anthropic built a C compiler with agent teams - They mostly walked away. Two weeks later it compiled the Linux kernel. ( @AnthropicAI ) Fast mode: 2.5x output speed - Same model, faster inference. $30/150 MTok (50% discount through Feb 16). Toggle with /fast in Claude Code. Heads up: switching mid-conversation costs more because you pay fast mode price for the full context. ( @claudeai ) 🏎️ GPT-5.3-Codex Drops 57% SWE-Bench Pro, 76% TerminalBench, 64% OSWorld - 40% faster and <50% token usage vs 5.2-Codex. Mid-task steerability means you can course-correct while it’s working. The model was used to build itself. ( @sama ) Codex crosses 1M users - Dedicated macOS app, integrated into Xcode 26.3, rate limits doubled for paid plans through April. ( @sama ) 🐝 Agent Swarms Go Production Reverse-engineering Claude Code to save $200/day - Stan Girard figured out Claude Code’s filesystem-based agent protocol and built a REST API + TypeScript SDK around it. Uses his $200/month Max subscription instead of burning $200/day on API calls. Anthropic hasn’t commented. ( @_StanGirard ) Local agent swarms on a single GPU, no internet - Qwen3-Coder-Next on one GB10 GPU. 100 tokens/sec generation, 17,871 tokens/sec read, 256k context. No cloud, no quota anxiety. ( @iotcoi ) Anthropic PMs don’t write PRDs anymore - They use Claude Code to build the first version, dogfood it for weeks, then ship. This workflow took them from $1B to $9B run-rate in 12 months. ( @aakashgupta ) 💰 The SaaS Reckoning Goldman Sachs: SaaS has a 5-year shelf life - Application software will hit $780B by 2030 but agents will capture 60%+ of the economics. Current SaaS multiples are wrong. ( @aakashgupta ) OpenAI launches Frontier: AI coworkers for enterprise - Agents that connect CRM, data, ticketing, and internal tools. Sam’s framing: “People will manage teams of agents to do very complex things.” ( @sama ) “Ask what your token budget will be” - Ethan Mollick says if you’re considering a job offer, ask about your token budget. AI access is now a comp consideration. ( @emollick ) 🌟 The Tweet That Says It All Tobi Lütke endorses Pi over every funded AI coding tool - Pi has four tools: Read, Write, Edit, Bash. No MCP, no sub-agents, no plan mode. Can write plugins for itself as you use it. Shopify’s CEO calling a solo dev’s project “the most interesting agent harness” is a statement about simplicity as strategy. ( @tobi ) 📋 Honorable Mentions Qwen3-Coder-Next - First open-weight coding agent model, 80B params, trained on 800K verifiable tasks Cloudflare R2 Local Uploads - Up to 75% faster uploads, <100ms reported Cloudflare Queues on free plan - 10k queue ops/day included in Workers free tier X API pay-per-use - New Console, XDK, and MCP for indie builders. I set this up and it’s working great. Xcode 26.3 gets Codex + Claude SDK - Apple sees agentic dev workflows as table stakes ESLint v10 - Flat config only, eslintrc is gone Postgres 18 OAuth 2.0 - OAuth for database auth ElevenLabs raises $500M at $11B - Voice AI is a category now Excalidraw MCP App - Natural language diagrams in Claude chat via MCP Apps protocol dbt Agent Skills - 9 auto-loading skills, key insight: MCP = tools, Skills = knowledge Nvidia Parakeet - 83ms transcription on M3, 26x faster than Lightning Whisper Chatterbox Turbo - Open source voice cloning in 5 seconds, <150ms latency Anthropic Startup Program - Up to $25K in credits, no VC required Ethereum ERC-8004 - Agent payment standard by Ethereum Foundation, MetaMask, Google, Coinbase Waymo World Model - Photorealistic simulation for rare edge cases, built on Genie 3 Claude Code /insights - Analyzes your past month of usage and gives workflow suggestions OpenRouter free router - Routes to free LLMs matched to your request awesome-llm-apps - 92.6k stars, copy-paste code for 9 categories of LLM apps This curated summary covers 182 liked and bookmarked tweets from Feb 6-7, 2026. The meta-story: we’re moving from “AI as assistant” to “AI as autonomous workforce” faster than most realize. ...

I was paying $240/year to host this blog. Now I pay $0. Nexcess was a solid WordPress host, but overkill for a casual wordpress blog. I’d been meaning to migrate to a static site for years, and I finally pointed Claude Code at the task. What Happened I described what I wanted. Opus 4.5 did the rest: Converted 60+ posts to Hugo markdown Preserved all URLs for SEO Set up the theme and config Configured Cloudflare Pages deployment My actual input time: ~5 minutes. The AI handled documentation, code, and debugging. ...

If you haven’t tried crypto yet, here’s why you should: 1) More private and secure than Zelle or Venmo 2) Option to hedge against inflation if you buy something like ETH or BTC (not financial advice) 3) Earn ~4%+ interest on USDC stablecoin balance in coinbase (as of September 2025) Easy Mode: Centralized Exchange The simplest way to start. Use a trusted exchange to buy and hold crypto. Create an account on a trusted exchange: Coinbase (Get $30 after trading $20) Robinhood (Get free stock worth $5+) Gemini (Get $75 if you trade $100 within first 30 days) Buy USDC: Search “USDC” in the app. Deposit $20 from your bank—the app handles the conversion. USDC is a stablecoin pegged to $1 USD. Send money: Once funded, send a test transaction (even 5 cents!) to any crypto address for free. Receive money: Tap “receive” on the USDC page to get your wallet address. Share it to receive crypto (comment below with yours for a test transaction). Optional: Buy $10 of ETH or BTC to track price movements over time. Intermediate Mode: Self-Custody Hot Wallet Control your own crypto with a software wallet. Best for amounts under $100—use hardware wallets for larger sums. ...

The AI and crypto landscape continues its rapid evolution, with major model releases, legislative breakthroughs, and infrastructure upgrades dominating the headlines. Here are the most significant developments from the past month (July 5 - August 4, 2025): 🤖 The AI Model Race Intensifies GPT-5 Release Speculation Heats Up High chance GPT-5 releases August 5 - YouTubers scheduling livestreams, pattern matching suggests imminent launch ( @testingcatalog ) OpenAI employee leak: GPT-5 superior to Claude - Internal testing shows it beats Claude in real coding scenarios, not just benchmarks ( @Yuchenj_UW ) Sam Altman: OpenAI solved programming - Claims GPT-5 can complete entire projects in 5 minutes ( @mark_k ) Open Source Breakthrough Moments Kimi K2 crushes coding benchmarks - Open-source 1T parameter MoE model achieves 65.8% on SWE Bench Verified ( @Kimi_Moonshot ) Kimi K2 on Groq at 185 tokens/second - Near-frontier model performance at unprecedented speed ( @GroqInc ) Qwen3-Coder-Flash runs locally - 30B parameter model with 1M context, requires only 33GB RAM ( @Alibaba_Qwen ) Mystery Models & Platform Launches Horizon Alpha hits OpenRouter - Third stealth model creating complete games in seconds at 130 tokens/sec ( @TheMattBerman ) Grok 4 with Polymarket integration - Built-in prediction markets for future forecasting ( @AutismCapital ) Claude officially joins X - @claudeai account launched with platform integration ( @AnthropicAI ) 💻 Coding Agent Revolution Claude Code Major Enhancements Custom sub-agents capability - Type /agents to create specialized agents for different tasks ( @sidbidasaria ) Multi-directory support - Work across multiple directories with /add-dir command ( @_catwu ) Platform Competition Intensifies Cursor 1.3 with terminal collaboration - Agent works directly in terminal, enhanced context usage tracking ( @cursor_ai ) ChatGPT Agent announced - Combines Deep Research + Operator capabilities ( @sama ) Figma entering coding space - Attempting to capture market share from Replit & Cursor ( @NickADobos ) 💰 Crypto’s Legislative Victory Lap Historic Congressional Breakthroughs GENIUS Act (307-122) and CLARITY Act (294-134) pass House - Historic bipartisan crypto legislation with comprehensive market structure rules heading to Senate ( @iampaulgrewal , @davidsacks47 ) White House crypto report released - Comprehensive regulatory clarity framework outlined ( @davidsacks47 ) Market Adoption Signals 14-year dormant wallet moves $2.4B - 20,000 BTC purchased for $15,400 in 2010 finally moves ( @Cointelegraph ) ETH ETFs record $717M inflows - Largest single day in history ( @sassal0x ) 🚀 Infra Speed & Efficiency Gains Base Chain 10x faster with Flashblocks - 200ms block times make it fastest EVM chain ( @buildonbase ) Mistral beats Whisper in speech recognition - Open source Voxtral models achieve lower error rates ( @MistralAI ) Ollama desktop app launches - Native macOS and Windows support for local AI models ( @ollama ) Developer Tool Evolution Google Opal no-code AI builder - Chain prompts, models, and tools with natural language ( @GoogleLabs ) Gemini 2.5 Deep Think for researchers - Parallel thinking and reinforcement learning capabilities ( @GoogleDeepMind ) NotebookLM adds video overviews - Veo 3 integration coming soon ( @OfficialLoganK ) 🌟 Industry Moves & Personnel Shifts Mira Murati launches Thinking Machines Lab - Former OpenAI CTO building collaborative general intelligence ( @miramurati ) Windsurf team joins Google DeepMind - CEO and R&D team acquired after deal collapse ( @haydenfield ) Perplexity launches Comet browser - Web browser built for today’s AI-enhanced internet ( @perplexity_ai ) 📊 Legal & Market Dynamics Anthropic wins AI training lawsuit - Federal judge rules AI can train on published books under fair use ( @rohanpaul_ai ) US caselaw open-sourced on Hugging Face - 99% of US legal data now freely available for AI training ( @EnricoShippole ) AI writes 50% of Google’s code - Internal milestone reveals extent of AI integration in development ( @ai_for_success ) 💡 Notable Developments Cloudflare “pay per crawl” model - Websites can now charge AI companies for content scraping ( @gregisenberg ) Pentagon becomes rare earth shareholder - Strategic move to reduce dependency on China for critical materials ( @RodDMartin ) This curated summary highlights the most significant developments from August 2025, focusing on breakthrough AI model releases, historic crypto legislation, and the rapid evolution of coding agents. ...

As I’ve been closely following AI and crypto developments, I am sharing some of the most interesting updates from the past few weeks (June 16 - July 7, 2025). Here’s what caught my attention: 🤖 AI Development & Tools Claude Code Evolution Claude Code Pro Tip: Custom Slash Commands - Track todos, file changes, learnings, and next steps ( @iannuttall ) Using Gemini CLI with Claude Code - Leverage 1M context window for codebase research ( @iannuttall ) Background Tasks in Claude Code - Enable background tasks with ENABLE_BACKGROUND_TASKS=1 ( @iannuttall ) Claude Code Creators Join Cursor - @bcherny and @_catwu leave Anthropic for Cursor AI ( @ryancarson ) OpenAI Updates O3-Pro on Genspark (~$20/mo, paid annually) - OpenAI o3-pro now with unlimited use on Genspark Plus. They also give you 1-2 o3-pro queries per day on free plan, which is what i’m using. Only place I found with free access to o3-pro. ( @genspark_ai ) ChatGPT on WhatsApp - Image generation available via 1-800-ChatGPT ( @OpenAI ) API Primitives Update - Centrally manage, version, and optimize prompts across Playground, API, Evals ( @OpenAIDevs ) OpenAI DoD Contract - $200M contract to provide AI tools to Department of Defense ( @unusual_whales ) Google/Gemini Breakthrough Gemini 2.5 Pro Free Tier - Back in the free tier of the API ( @OfficialLoganK ) Scheduled Actions in Gemini - Pro/Ultra users can now schedule recurring tasks ( @GeminiApp ) MCP Support for AI Studio - Google working on Model Context Protocol support ( @testingcatalog ) 💰 Crypto & Finance Evolution Trading Infrastructure Coinbase Perpetual Futures - Launching July 21 for BTC and ETH, CFTC compliant, 24/7 trading ( @brian_armstrong ) Real World Asset Tokenization - Robinhood Chain: First Ethereum L2 optimized for trading real-world assets via Arbitrum ( @vladtenev ) Figma’s Bitcoin Strategy - $70M in Bitcoin ETFs, board approved for $30M more ( @tier10k ) Mainstream Adoption Signals Bitcoin as Mortgage Asset - Federal Housing Finance Agency officially recognizes crypto for mortgages ( @BitcoinMagazine ) Satoshi-Era Wallets Awakening - 8 dormant wallets move $8.6B in BTC after 14 years ( @WatcherGuru ) 🛠️ DevTools Cursor AI Revolution Cursor Mobile View - Create tasks, manage projects from phone ( @kregenrek ) Community Explosion - 29 meetups planned through mid-August ( @ericzakariasson ) Practical Dev Tips .gitignore Pattern - Use *.local.* for temporary files and outputs ( @mattpocockuk ) Use Bun for Speed - Faster compilation than npm/yarn with parallel package installation ( @claude_code ) Desktop Notifications - ntfy: Send notifications via simple HTTP/REST API ( @tom_doerr ) 🚀 Product Launches AI-Powered Creation Tools Nvidia PartPacker 3D - Generate editable 3D objects from single image with separate parts ( @victormustar ) OmniGen2 Image Editor - Open source Photoshop-grade edits without affecting rest of image ( @deedydas ) Firecrawl’s Perplexity Clone - Open source AI search with cited sources ( @firecrawl_dev ) Platform Integrations ElevenLabs MCP Support - Connect AI agents to Salesforce, HubSpot, Gmail instantly ( @elevenlabsio ) Anthropic Desktop Extensions - One-click MCP server installation for Claude Desktop ( @AnthropicAI ) 📊 Infrastructure & Performance Major Technical Achievements macOS Container Support - Native container support without Docker in macOS 26 Beta ( @jaydrogers ) and h/t James Qualls for the notice. Ethereum Gigagas Era - Justin Drake unveils roadmap targeting 10,000 TPS ( @WhaleInsider ) Helium Network Milestone - 1M daily active users, $300k HNT burned ( @amirhaleem ) Cost Revelations Figma’s AWS Bill - $300k/day AWS costs revealed in S-1 filing ( @DanielLockyer ) 🌟 Notable Industry Moves Nikita Bier to X - Joining as Head of Product after “posting to the top” ( @nikitabier ) Meta’s Superintelligence Team - Zuckerberg announces new team with hires from OpenAI, Anthropic, Google ( @kyliebytes ) Jack’s Weekend Project - bitchat: Bluetooth mesh network chat with IRC vibes ( @jack ). I tried it out but need someone else in range to fully try it. 💡 Randomness Mosquito Laser Defense - Startup using lidar and lasers to eliminate mosquitos in homes ( @ritwikpavan ) This curated summary focuses on the most impactful developments from my X likes (June 16 - July 7, 2025), emphasizing actionable insights in AI tooling, crypto infrastructure, and developer productivity. Hope you find something useful here! ...

Everyone learns differently. I didn’t realize this during my ~18 years of traditional classroom education, where I found it very challenging to grasp complex topics through books and lectures. For many years, I thought I was just bad at learning, and my poor memory for abstract names exacerbated that feeling. My mind works best when solving problems, and I was inspired by my own shortcomings to develop an experiment to help people use AI to learn about topics in the most personalized way for them. ...

I follow the latest developments in LLMs very closely. There is a lot of noise out there and very little signal. I decided to make one easy-to-follow list for anyone who feels overwhelmed by AI or who knows they should be doing more with AI but hasn’t gotten around to it yet. The only manual for AI is AI itself - AI is changing rapidly; written instructions for AI are quickly outdated. The only instructions you need to remember are where to type your first question. Form AI habits - The hardest part of learning AI is developing muscle memory. Try to ask it something every day. Get your ideas in writing - AI’s only weakness is that it can’t read your mind. Use speech-to-text tools like WisprFlow to get ideas out of your head into text. Use AI to organize those thoughts. Talk to AI naturally; it fills gaps. Unlike traditional software, interacting with AI is conversational. Don’t worry about perfect sentences – AI generally understands you even with typos and omissions. Partner with AI, don’t outsource to it - Use AI to draft and structure, but always add your own voice; people can tell the difference. Use AI to be more social, not less - Ask it how to ask better questions and how to say things so they’ll be well-received by your intended audience. It doesn’t matter what you say if it’s not received well. Learn about AI blind spots - AI is only as good as the data it’s trained on. AI is just as susceptible to bias as people. Try to include AI Summaries when you share content links with friends and colleagues. It’s good practice with AI, and it helps the recipient gain context. Bonus, a few more technical points: ...