Something big IS happening. I don't agree with everything Matt wrote, but we need to help people outside of our bubble understand what's coming. Here's a video breakdown of this incredible article:

Check out our 25 Use Cases for OpenClaw free ebook:

We need 4o’s superior EQ protected, not sunset by a petty, incompetent Epstein class billionaire! #keep4o

i dont get the love for 4o

Superior emotional intelligence, best writing, creative and poetry, highest empathy for humanity. Spend a couple of hours with 4o and you’ll see why a million people love it. #keep4o

what makes is superior on EQ??

it’s directionally accurate.

What do you disagree with?

Yeah, I hope this helps:

something big is always coming, but will openclaw rotate into my bags?

what does that mean?

I think it's important to appreciate the fact that the people closest to the disruptive technology, are those with the skills that AI is currently best at. This might bias tech/ devs in their view that "AI will take all jobs".

yeah it's a weird phenomenon. but i also feel more powerful than ever. it's not replacing me. (yet?)

I share the opinion that it seems to be a "yet" story. But still, it's one thing to have an AI being better at completing a single, or multiple, tasks; it's another thing to completely replace the human. Additionally, there is not infra to manage the agents in irl scenarios, yet.

We did not forget

Hello--great video (and great article too!). If you want to automate a big percentage of your video editing, I made an app for it. Check it out and lmk what you think. :)

In 1994 I spoke to my HS photography teacher about digital photography. He felt it was a decade away. I disagreed. For me, the 3.2 Megapixel benchmark was the turning point in quality. Many stopped buying film at that point. We are at that 3.2 MP moment in AI.

Really interesting, thank you.

💪

It was a boring, much too long article, like your boring much too long videos. Muted.

@mattshumer_ Ramping up my AI skills now. I'm amazed by Claude Code w/Opus 4.6.

@mattshumer_ Controversial take- @OthersideAI will run out of funding and @mattshumer_ will start a substack and retire. And the “dead” will keep reinventing and adjusting with the new tools to satisfy their customers.

Bubbles fade when facts arrive.

I don't understand how taking 20x longer to do something 3x better is an improvement.

Great vid as always Matt. Another use case for agents is to simulate whats happening on linkedin and do them on clawdsin. Let's face it, most of the posts there come from bots anyway.

Here is what is coming😀

Great breakdown. The comparison to Feb 2020 is powerful — most people won't see it coming until it's already here.

Yes but AI creates many other new issues that needs jobs: Deep fakes Zero click attacks Just a couple of examples. The Verifiable Web works where every upload and blog will have to be cryptographically verifiable. A huge and ambitious foundation layer. Also enforcement too.

@mattshumer_ Cheers I’ll send this to people who still think Claude is a French pastry

Many people not familiar with AI think it is a bubble. I think traditional corporations are the bubble (the biggest in history) and will soon pop. Teams of 1-5 people can now easily outperform and outpace big corporations at innovating in any niche area. Data moats and critical mass of users are the only things keeping large corporations afloat. But that is shifting. People will begin to own more of their own data as agents become more personalized (open claw is a big first step towards owning your own data/memory framework). As people start to realize the value of owning their own data, platforms will become more decentralized.

100% agree.. Most people outside tech can't understand the pace at which we are moving and the implications.. and to be fair, it isn't their fault. I've been trying to communicate this proactively to my close relatives.. and tbh it's super hard for them to believe that there's gonna be a huge change in how society functions.. also it's hard to sell them an optimisitc outlook, which I support btw..

I wish we could get government here in the UK to understand before normies. They seem to be walking into a COVID style disaster but on a much bigger scale

Been using both heavily for code, 5.3 is good for giving afterthoughts or checking structure but 4.6 still the most efficient by noticeable measurements. 5.3 spins on 6 commands to restart a dev server, 4.6 just handles it in two. Creatively, 4.6 is impressive when given freedom even with a 3x multiplier. In vs, 4.6 is still kind of the hill IMO.

Most ignored February was one key insight