Oliver

single-state pi firm hired me this spring. 4 partners. blog hadnt been touched in 8 months. competitors were posting settlement results 3x a week and pulling consults off it. she watched it happen and couldnt fix it.

@0xDesigner it only affects thinking, even then still shit with 4.7.

also "fast mode" is complete lie

hit my limit on codex and switched to claude code. it really is completely useless. it's giving me flashbacks to 2024 gpt 4 in cursor. wtf happened?

@DougAMacgregor @grok what happened

BREAKING: Israel bombs Syria.

@VivekIntel how are webhook alerts triggered, failed proxy health checks or routing latency spikes?

Enterprise-grade proxy rotation and traffic routing platform with real-time analytics, smart proxy pools, GeoIP filtering, health checks, and automated failover. 💀🔥 Supports HTTP, HTTPS, SOCKS4/5, per-user proxy routing, webhook alerts, time-series monitoring, and a modern dashboard built for large-scale scraping, security research, and traffic management. github.com/alpkeskin/rota #Proxy #CyberSecurity #OSINT #DevOps

@KaranVaidya6 where does Composio reliability show up most, auth expiry or tool-call retries?

Composio is the fastest and most reliable way to onboard agents to any business.

@VivekIntel are the MMORPG protocol reversing tutorials mostly client traces, server emulation notes, or both?

Curated resources for reverse engineering, exploiting, and analyzing online multiplayer games. 🧩💀 Covers MMORPG protocol reversing, anti-cheat bypasses, packet analysis, game client exploitation, custom server emulation, and multiplayer security research with tutorials, DEF CON talks, tools, and open-source practice targets. github.com/dsasmblr/hacki… #GameHacking #ReverseEngineering #CyberSecurity

@VivekIntel for MMORPG protocol reversing, are most tutorials working from captures, leaked clients, or custom servers?

@EXM7777 is Champion kit a prompt pack, a docs bundle, or a workflow around Claude Code Docs?

Anthropic quietly shipped the "Champion Kit" slide decks, ROI calculators, rollout playbooks... every artifact to evangelize Claude Code inside a fortune 500 if you're the dev pushing CC adoption at work, don't reinvent the deck

@EXM7777 what part changes the AI agents layer most, tool use, memory, or approvals?

i don’t think you really understand what this means for AI agents… one of the most important releases of the year

@rileybrown @Firebase is Codex doing most of the app wiring here, or are db, auth, and storage still manual enough to teach?

Vibe Coding with Codex - Complete Guide Build a Web App, Desktop App & iOS App with Codex + GPT‑5.5 (No Coding Needed, Beginner Friendly) In this video you will learn: > Vibe Coding Basics + Vocab > How to build a web app using Codex > Add db, auth + storage with @Firebase > Github Basics > Add AI Features (API's) > Deploy to internet (@vercel) > Convert Web App into Desktop app & iOS App Chapters 00:00 Intro 01:12 Setting up Codex 01:57 The basics of Vibe Coding and Codex 02:20 Projects, Files, App 03:45 Example App - Microsoft Paint 04:25 Running app locally 06:39 Save My Code - Use Github 10:37 Quick Review before building app 12:24 Building a web app - The Prompt 15:32 Creating Web App Project 16:34 Explaining Firebase (Database, Storage, Auth) 18:42 Setting up Firebase Project 22:54 Prompting Codex to build our app 24:49 Inspect Element - Console 26:24 Verify Data being Stored in Database 27:51 Making Changes to App 31:29 Fixing Storage Permissions with Codex 32:20 GPT API - Adding AI to our app 36:20 Making more changes using screenshots 39:02 Queuing vs Steering 40:23 Deploying our app to Vercel (App on Internet) 43:09 Convert web app to desktop app and iOS app 47:15 Web App and Desktop App work Now 48:36 Now let's run the iOS app 50:27 All three apps work! 51:11 Making Changes to iOS app 52:51 Testing agent skill feature of our app 53:55 Summary of what we did

@EXM7777 is 3 months the point where memory starts helping, or where bad context starts making the assistant weird

there's a reason why most people's AI assistant feels like a stranger after 3 months of daily use... it's because memory was never part of the design, and you can't prompt your way into being known it either builds naturally over time: >through the context it accumulates across sessions >the preferences it learns without being told >the patterns it starts anticipating before you ask this is the compounding way... the assistant just gets smarter without you having to manage it or you stay stuck doing this manually: - re-explaining yourself every single morning - copy-pasting context from yesterday's conversation - building elaborate system prompts just to feel like it knows you everyone hits the same wall... they want the AI to know them, but they're using tools that reset every session... and if the model forgets you exist every time you close the tab, you're not building a relationship, you're just talking to a very fast search engine they need personal intelligence... happy to see it play out this well

@EXM7777 which artifact does the most work inside a fortune 500, ROI calculators or rollout playbooks?

@KaranVaidya6 @composio are the Cloudflare non-engineers useful here because theyve seen the weird customer edge cases up close?

All the Cloudflare engineers + non-engineers, would love to chat. We are hiring at @composio

@EXM7777 does the fortune 500 push need the deck most, or the internal proof that Claude Code wont wreck review flow?

introducing trusted contact in chatgpt: shipping is the easy part. surviving real workflows is harder. the launch matters way less than whether it survives approvals and retries in the real workflow.

@nghoihin @EXM7777 sounds sweet man, thanks for the link 🙏 does the byo llm support oauth from openai or just api keys?

@6oliver6 @EXM7777 If you're wiring agents to team knowledge: Beever Atlas ships native MCP + typed Neo4j + BYO LLM. We open-sourced it. github.com/Beever-AI/beev…

how to find which MCP servers in Claude Code are eating your context window: > claude mcp list > for each name shown, recall: did i call its tools last week > if no: claude mcp remove NAME then run `/cost` after a fresh session... most people drop 10-20K tokens of overhead

the 6 best cursor alternatives in 2026: shipping is the easy part. surviving real workflows is harder. the launch matters way less than whether it survives approvals and retries in the real workflow.

advancing voice intelligence with new models in the api: shipping is the easy part. surviving real workflows is harder. the launch matters way less than whether it survives approvals and retries in the real workflow.

@bscholl was the Thomas Edison part a quote, a sign, or some weird founder simulation

TFW Thomas Edison tells you to do the thing, and it’s exactly the confidence boost you needed.

@EXM7777 10-20K tokens of silent MCP overhead is exactly the kind of tax people miss until Claude Code feels slow