AI Engineer | Working with Agents I build AI explainers with code deep dives on YouTube

Pinned Tweet
I can officially retire from X now. Sam Altman just quoted my tweet. 🔥
12
63
8,335
GEMINI 4 IS ENTERING ITS FINAL STRETCH. A suspected Gemini 4 Pro checkpoint is being tested on Arena. The next model war is getting closer.
1
78
OpenAI may be building agents that never stop working. A leaked model ID 'gpt-6-astra-aeon' is fueling speculation about “Aeon,” a persistent agent system designed for long-running tasks that keep going in the background. Close the laptop. It keeps working.
4
5
495
OpenAI may be changing how developers build and host AI apps. Leaked references point to three new platform tiers: → Free → Prototype → Accelerate: reportedly starting at $50 The plans could bring app building and hosting deeper into OpenAI’s developer ecosystem.
1
3
280
Claude Code just fixed one of its most annoying problems. Hit the 5-hour limit mid-task? It no longer cuts off mid-edit. It now tries to find a clean stopping point and uses a small weekly allowance to wrap up.
Claude Code will now try to find a graceful stopping point when you hit your 5-hour limit mid-task, instead of cutting off mid-edit. It gets a small, fixed allowance pulled from your weekly limit to wrap up what it can.
1
2
192
Tibo just said Codex, ChatGPT usage limits are being reset for all paid users. I still have 5% left. Reset is supposed to hit on the 28th. Did anyone actually get the reset… or is it still rolling out?
13
32
5,501
OpenAI, Codex ChatGPT is up and working now
We are aware that codex is down and are working hard to bring back normal service.
6
385
Tibo will reset usage limits for all paid users accross Codex and ChatGPT Work. The question is whether it will be a banked reset or what?
o yes… we’re back in action and we’ll reset usage limits for all paid users across codex and ChatGPT work sorry about the brief disruption! (and yes we have a special spare codex when things are down to help us out)
1
18
1,185
Gemini 4 Pro checkpoints are already being tested. Multiple Arena sightings under the “gemini-3.8-flash” label have produced outputs far beyond any normal Flash model, complex SVGs, and interactive games. Internal codename circulating: Argon
31
1,556
GPT-6 Luna just got exposed. GPT-6 Luna at max: 37 GPT-5.6 Luna at max: 37 A whole new generation. The exact same score. Meanwhile, Qwen, Grok, and GLM are hitting 44–45. OpenAI slashed the price in half. But Luna hasn't gotten any smarter.
16
1
70
5,356
GPT-6.1 Astra rumors are heating up 👀 OpenAI may be preparing another major model upgrade just days before DevDay. The big question: is GPT-6.1 Astra a more powerful version of Astra, or an entirely new tier?
10
14
205
13,125
Gemini 4 Pro is routing through "Gemini 3.8 flash" in Arena AI SVG image of Pelican
6
1
51
1,895
Did Google just leak Gemini 4 Pro? A mystery model named "gemini-3.8-flash" is in test in LMArena right now, but the stats prove it's a hidden flagship: 256K outputs (4x the actual Flash cap) One-shot 3D worlds & flawless complex code Project Argon in disguise?
4
3
105
5,025
Google’s next image model is already testing under the codename “Spicy Mayo.” Early Arena results show big jumps in speed, texture detail, and text rendering over Nano Banana 2. Endpoints for nano-banana-2.5-flash are also showing up in Google Flow.
1
21
880
GPT-6 Cyber may be unveiled at OpenAI DevDay. Reports point to a specialized cybersecurity model designed to detect vulnerabilities and help automate security workflows. OpenAI may also introduce a product to help organizations deploy it securely.
1
21
1,671
Gemini 4 Pro leaks are getting harder to ignore. A suspected internal checkpoint, “Argon,” was spotted with a 256K output limit. Now there are reports of a newer “Argon-d” checkpoint being tested. If these are real, Gemini 4 is already deep in iteration.
6
3
116
5,377
OpenAI may be building an entire platform for autonomous agents. A leaked feature called Managed Agents points to tools for building, hosting, and managing agents directly on the OpenAI Platform. Not just chatbots. Agents that can run multi-step workflows in managed environments.
3
276
OpenAI DevDay leaks: A $500/month ChatGPT Pro Max tier has appeared in unreleased configs. The pitch? “Fastest Work and Codex.” DevDay is going to be interesting.
9
12
1,529
OpenAI is dropping something on Tuesday Guess what?
Been a bit quiet here because internal Slack has been hilarious lately and because we are all locked in on DevDay. Tuesday will be fun.
2
1
21
1,385
OpenAI is testing a $500/month ChatGPT plan. Codename: Pro Max. Current top tier is already $200 and they still had to pause new sign-ups because of demand. Now they’re preparing something 2.5× more expensive for the heaviest users.
1
10
284
Grok 4.7 is now getting beaten by even Muse, Hy4, and even Gemini 3.8 Flash. On Agent Arena it ranks #16 Claude and GPT-5 are posting double-digit gains. Meanwhile Grok is stuck behind smaller models that cost a fraction of the hype.
5
8
508