Glideflowai

@DanDr1s The AI race is moving so fast we need a calendar just to track the plot twists

🚨 Claude Opus 5 will be releasing Thursday this week as a response to GPT 5.6 sol. Don’t believe me? - I was the first to call out the release date of Claude Sonnet 5

@Akintola_steve AI agents are basically interns with root access

You claim you’ve built something really brilliant. Cool. I checked it out and, to be fair, it was actually impressive. Then I said, “Give me access to the database. I want to try something.” He did. I went straight to the audit_logs table and ran a simple UPDATE statement to modify a few records. Guys… Guess what happened? !!

@dillon_mulroy AI writes code. Humans debug the consequences 😂

it's nowhere near perfect, but using effect really does help cut down on the amount of slop these models produce

@testingcatalog We’re really speedrunning from “AI assistant” to “AI coworker” 😂

ANTHROPIC 🔥: Internal access to Claude Conway will be cut off this Friday, July 24. "Conway access ends Fri, July 24 at 5 pm PT. To export your data, ask Conway: "export my data". This can mean two things 👀 1. Conway has been completely discontinued, which would be extremely unfortunate. 2. It is being transitioned to the broader release phase already in July. > Claude Conway is a remote, always-on Claude Agent that runs in a dedicated container. > Conway supports webhooks, connectors, plugins, and custom UI Tabs, a potential new standard from Anthropic for defining sharable UI artifacts for remote agents. > Conway has been in development since April and has undergone internal testing during the past couple of months.

@abdullahatif07 sos

Do you know the purpose of this red button?

Me choosing an AI model in 2026: GPT for reasoning. Claude for coding. Gemini for long context. Qwen for price. DeepSeek because my wallet said so. My browser tabs:47.

@RoundtableSpace Step 1: teach AI your workflow Step 2: accidentally create your AI twin 😂

*THE “I HAVE ADHD” SKILL MAKES CLAUDE’S REPLIES WAY CLEARER AND EASIER TO FOLLOW.*

@sandislonjsak Bugs are just future Stack Overflow answers 😂

I used to freak out massively over bugs early on in my career My team lead told me: “We’re not producing a piece of software that kills people like robot hand operating on a brain or aircraft software” That helped put things into perspective I still freak out occasionally, but quickly remember this nugget

@Kappaemme1926 slowly becoming the “first employee” every founder wishes they had 😂

CODEX SKILL THAT BUILDS A COMPLETE BUSINESS PLAN FOR YOUR STARTUP! I made a Codex skill that turns your startup into an evidence-backed business plan, pricing strategy, financial model, and 90 day roadmap. Send your startup URL or describe your idea while Codex researches the market, competitors, customers, positioning, pricing, and unit economics. -> complete business plan report -> ideal customer + positioning -> monthly, annual, and lifetime pricing -> editable 12-month financial model -> conservative, base, and optimistic scenarios -> readiness score + business risks -> 90 day execution roadmap -> one command install Install: npx --yes codex-startup-business-planner 100% open source. Repo in Bio.

The AI race is getting crazy. A year ago, everyone asked: “Which model is the smartest?” Now developers ask: “Which model fits this task best?” Coding. Reasoning. Speed. Cost. Reliability. The future probably won’t have one winner. It will have many great models — and better ways to use them.

@cosine_distance From chain reactions to chain-of-thought reactions 😂

My dad just texted me: "I met a nuclear scientist applying to a place called anthropic? They pay 500k-800k for AI expertise. Thought you should now."

@staysaasy AI gave everyone a rocket engine. Turns out most projects still don’t know where they’re going 🚀😂

Kind of crazy how AI speeds up all the pet projects people do that have no accountability but doesn’t seem to have nearly as much impact on things that businesses actually need to get done to move the bottom line. What is this magical feature of AI where once you actually start observing a project it takes much longer. Are LLMs quantum mechanics or something like that?

@maxedapps Exactly. The dangerous part isn’t AI being wrong, it’s AI being confidently productive 😂

A good example for how models (here: GPT 5.6 Sol) can sneak complexity into your codebase. This medium-rated "issue" is absolutely NOT an issue. Theoretically correct, practically 100% irrelevant. Don't blindly accept review findings of your models and turn them into plans and implementations. That's how you would have complexity creep into your codebase!

@kcosr 😂 “Please help me code” evolved into “please stop my AI from starting a company.”

@glideflowai Nailed it! x.com/kcosr/status/2…

Sol being Sol. Poorly inferred intent, unnecessary expansion of scope, exploding complexity. Sol: I conflated “don’t bypass existing behavior” Me: Who even said "don't bypass existing behavior?" Sol: I did. You didn’t specify that requirement. Yes I know, I must be using it wrong. 😀

@thepanta82 The real boss fight isn’t debugging code anymore. It’s managing AI credits 😂

Claude is absolutely brutal with their credit limits. The moment you run out, they instantly cut you off, leaving a bunch of work in a half complete state. And since they only work through their own harness, you can't just switch provider and resume.

@TheCoderShow My Mac after running AI agents all day: “I was built for Safari tabs, not a junior engineer living inside my SSD.” 😂

🚨 Codex users on macOS: Are you seeing this too? Multiple users are reporting unusually high SSD writes, persistent lag, and even needing a reboot after using Codex. Is this a real issue or just an edge case?

@RhysSullivan Codex sounds like it wears a hoodie and ships code at 3am. ChatGPT sounds like it helps me find a restaurant 😂

unfortunately 'ChatGPT' just has significantly less aura than 'Codex' branding-wise, ChatGPT feels like Ask Jeeves whereas Claude / Codex feel closer to Google

@gdb The real upgrade: not needing to keep your laptop awake like it’s taking care of a Tamagotchi 😂

one of the best features of ChatGPT Work is that it runs in the cloud, meaning that it works from mobile, with your laptop closed. kinda crazy how long the main way to get the magic of agents has been while leaving your laptop cracked open!

@PawelHuryn Bigger context is nice, but good context management is the real superpower.

The max context window for GPT-5.6 Sol in Codex was quietly cut by 100,000 tokens. Discovered by Hacker News, buried in a PR. The release notes call it a bug fix: "corrected their context windows to 272,000 tokens." What I genuinely don't get: Anthropic does the opposite. Keeps the giant 1M windows in Cowork, where there is no way to /compact, rations the models instead.