ChatGPT 5.6 is the greatest technology I've ever used I've been testing it in early access and I can confidently say it's as smart as Fable for a fraction of the price It's harness is also 100x better Here's everything you need to know about getting the most out... of ChatGPT 5.6show more

To the people who are crying because I use 'best model ever' quite often: I'm not testing cheeseburgers here. I'm not testing a product that's been solved for hundreds of years I'm testing an emerging, hyperscaling technology that is getting trillions of dollars invested into it You should be more surprised if I ever review a model and DON'T say it's the greatest ever Why would they release a model that was worse than the previous version? If OpenAI released ChatGPT 5.6 and it was worse than 5.5, the global economy would crash. It would prove the trillions of dollars being invested into AI was a waste and we hit a wall "Greatest tech ever" should be the expectation for every model release. Anything less is a massive disappointment and horrendous waste of money What should be more surprising? If Opus 5 is better than Opus 4.8 or worse than Opus 4.8? As for why almost EVERY review I do is highly complementary: I don't review models I don't like. I don't use models for 5 minutes then film videos. I use models for hours if not days, build tons of tooling and products, then tell you what I think. I'm not going to use crap models for hours and waste my time. Why would I use anything other than the best? It's the reason you've never seen me release a video on a Meta model or 99% of the Chinese models. They're not good. Why would I spend hours using them? I'm only going to use the best models for enough time to give you an educated review If I ever release a video for a model and don't call it incredible, then that company must have put a gun to my head because I'm not going to spend hours using a technology that doesn't impress me If any of this bothers you, I got good news, the unfollow button is in the top right. I highly encourage you to use it. Otherwise, I hope you enjoy these videos! They come 100% out of passion and wanting to help educate others. I've taken maybe 3 sponsorships the last 5 years. I'm not going to be one of those phonies that put 'promoted post' on 50% of their content and put cringe lame reaction gifs under corporate slop captions and launch videos. This content comes purely from the love of technology, and I hope that shines through!

This guy! Everything is always the greatest with him. Makes one question the credibility of what he is saying

When everything you try is the “greatest technology .. ever” nothing is the greatest.

Another day, another “GREATEST TECH EVER MADE” post from the Grift King.

The hyperbole again…?

the amount of times @AlexFinn says "model xxxxis the greatest technology I've ever used" is insane uses this line every other day

Bro, you said the same thing about openclaw 6 months ago.

Love your content, but you literally say the same thing about every new model release

Do you trust it with send access to your emails?

they hit it out of the ball park with this one

ever heard of fable 5?

@AlexFinn haha yeah, calling things "best ever" is kinda the norm in tech. new stuff drops every day, gotta hype it up sometimes!

My Current Setup (for Alex Finn): I have a highly customized Claude Code workflow with: - A `.claude/` folder containing 4 specialized agents (designer, engineer, investigator, marketer) - 7 custom rules + several skills and writing styles - PowerShell hooks for guard-write, git context injection, and verification - A structured pre-commit review process with DECISION-LOG and retrospectives - Heavy focus on tech debt refactoring, brand voice consistency, and iOS/Capacitor-specific edge cases - Currently using Claude Opus 4.8 on Max 5x plan (~$100/mo) My work is precision-oriented rather than high-volume feature building. Question: Given the above setup, where exactly would GPT-5.6 Codex (especially in Cursor) give me a meaningful advantage? I keep hearing very strong praise for 5.6, but I’m struggling to see where it would clearly outperform my current Claude workflow — especially regarding rule adherence, multi-agent coordination, and consistency on tech debt work. Would love to hear your take.

Fable 5 is still better for me

@AlexFinn sounds awesome! is the harness really a big improvement? i might have to try it out myself

@grok how many times has Alex said something is the greatest technology ever?

How are you feeling about X algo these days?

Everyone is saying @grok extremely close if not as good in a lot of aspects, for better cost/usage limits. Have you used it?

Alex Finn’s strong praise for GPT-5.6 suggests it is delivering real improvements in reasoning and usability. If it can match Fable level performance at a much lower price while offering a significantly better experience, many users and teams may start shifting their workloads. The real test will be consistency across different types of professional work over the next few weeks.

Ai is beautiful

When will US Plus users get access?

AI progress is truly accelerating.

FREE MASTERCLASS ON NEW CHATGPT 5.6

fable brains without the fable bill might be the best trade of the month tbh

Yeah your AI is loosing its luster. Tone it down or we stop watching 😁

The competition at the frontier is getting really interesting. Better performance at lower cost is a win for everyone.

does anyone use claude and chatgpt? I feel like having two subscriptions is a bit much

You definitely did NOT get early access and you definitely did NOT say it was awesome while testing early (iykyk)

Can't believe anyone on earth would trust this company with their data.

I'm going to use GPT 5.6 now.

Has anyone heard from Gary Marcus lately?

But for us it is not free and hard to use as a student

It's great, really useful in my daily life and work