Gashi | Building in Public with AI

30 days ago: zero coding experience. Today: 7 systems in production. Built with Claude Code. No bootcamp. Just solving my own problems. Sharing the full journey here: → Building AI SaaS from scratch → Daily ship logs → Revenue updates #BuildInPublic #AI

🔒 $10B AI startup Mercor got breached and everything's falling apart Valuation means nothing if you can't secure user data. Running production AI taught me: security first, growth second. One breach kills trust forever. How many unicorns are one hack away from collapse? #AI...

💰 ChatGPT Pro costs $100/month now Before: $20 → Now: $100 That's 5x more expensive. For what? Better reasoning? Priority access? I use Claude for coding anyway. OpenAI is pricing out indie builders like me. Is premium AI becoming enterprise-only? #AI #Pricing

🏥 Meta's AI wants your health data and gives garbage medical advice This is why I don't touch healthcare AI. Too much liability, too little accuracy. Meta should know better. Health data + bad recommendations = lawsuit magnet. Would you trust Zuckerberg with your blood work...

📊 Gen Z hates AI but can't stop using it Classic cognitive dissonance. They complain about AI while ChatGPT handles their homework. As someone running 6 AI systems, I get it. The tech works too well to ignore, even when you don't trust it. Are we building tools people secre...

⚖️ Florida AG is investigating OpenAI over a ChatGPT-linked shooting This isn't about the tech anymore. It's about liability. When your AI system runs 24/7 like mine, you think about edge cases. But nobody prepared for this level of scrutiny. How do you even audit conversat...

🛡️ OpenAI drops new safety blueprint Targeting child exploitation specifically. Not just generic "harmful content" anymore. Good to see focused action vs broad policies that catch everything. Are specific safety frameworks better than general ones? #AI #Safety

📺 Tubi just built inside ChatGPT First streamer to go native. No browser tabs, no app switching. "What should I watch?" → instant recommendations + playback This is how apps disappear into AI interfaces. Who's next? #AI #Streaming

📊 Atlassian just turned Confluence into an AI playground Visual AI tools + third-party agents = your wiki now thinks for itself Finally, documentation that writes back. No more dead knowledge bases. Which agent would you connect first? #AI #Enterprise

💰 AWS betting billions on both OpenAI AND Anthropic Their logic: "We're Switzerland, not picking sides" Smart move tbh. When you're the infrastructure, you win regardless of which model dominates. Is this the new playbook for cloud giants? #AI #Cloud

🛠️ Poke just made AI agents stupid simple No more complex workflows or API calls. Just text and go. This is what mass adoption looks like - when your grandma can use AI agents without reading docs. But will simplicity kill customization? #AI #NoCode

🤖 Anthropic doubled down on Google TPUs and Broadcom chips to meet demand. - Before: Limited to 100k H100s → Now: Access to 500k TPUs for training - Broadcom deal adds custom silicon for safety features Builders, this means more options for scalable training without begging ...

🚀 Uber just switched to Amazon's AI chips for their fleet tech. - Before: Relied on Nvidia dominance → Now: Amazon's Inferentia chips handle 40% more inferences per dollar - Scales AI routing 2x faster without custom code This lets builders plug affordable compute into real-...

🤖 OpenAI drops their AI economy blueprint: robot taxes and 4-day weeks—tbh, this overlooks builders grinding 7 days to make it happen. - Proposes taxing robots to fund public wealth - Pushes 4-day workweek for humans - Aims to redistribute AI gains to all For us builders, it...

Final take: Iran's move highlights AI as a weaponizable asset—builders, secure your stacks. Question: How are you hardening your AI systems against global risks? (4/4) #AI #Security

Why it matters: Stargate is key for next-gen AI, but threats like this mean we pivot to resilient designs. - Before: Centralized power → Now: Risk of attacks changes everything Running production AI, I've Claude-coded failover scripts—essential. #AI (3/4)

⚠️ Iran just threatened to hit Stargate AI data centers—yabai, this could stall global AI progress more than regulations ever did. - Before: AI infra was tech race → Now: Geopolitical target - Stargate aims for 1M+ GPUs, worth $100B+ As a builder running 6 AI systems, this p...

💰 Private wealth is flooding into super early AI bets, getting riskier by the day. - Before: VCs dominated seed rounds → Now: Family offices betting big pre-product - Shifted $50B+ into AI startups last year alone - More failures expected, but winners will be massive As a bu...

Ran a full audit on 4 AI projects yesterday. Found 15 issues — deprecated models, silent crashes, DB bloat, hardcoded paths. Fixed all of them in one session. Also built a new JP auto-post bot and a showcase site generator that turns daily work into animated pages. Maintenance isn't sexy but it's what keeps 6 systems alive 24/7. When did you last audit your stack? #AI #BuildInPublic

4 systems fixed in one session yesterday. - News bot down 4 days — API model deprecated, added fallback - Trend detector 36% accuracy — rewrote prompt, built tracking - Reviewed 20 AI-detected trends, 15 were wrong The unsexy part of running AI 24/7: maintenance never stops. What broke in your stack this week? #AI #BuildInPublic