Maximilian @maxedapps

I teach AI & coding: https://t.co/Pnlz9dslyF, do YouTube & build stuff like https://t.co/WQDO2d322k. Find me on https://t.co/OdF5RtTWHj. maximilian-schwarzmueller.com Joined August 2015

Tweets

8K
Followers

69K
Following

189
Likes

11K

Maximilian @maxedapps

16 hours ago

@amir0ffx Lol. Always complain, I guess? Of course it was easy to work around. Even nicer to not have to do it.

0 0 0 54 0

View Details

We're adding support for AGENTS.md to Claude Code. Starting today in version 2.1.277, if there is no CLAUDE.md in a folder, Claude will check for and use AGENTS.md. You can toggle this behavior in /config.

2K 3K 30K 4.7M 6K

4 2 147 34K 33

View Details

Maximilian @maxedapps

a day ago

@samgoodwin89 That’s pretty cool!

0 0 1 316 0

View Details

Maximilian @maxedapps

a day ago

@Ylodi @agents Symlink? :)

2 0 0 247 0

View Details

Maximilian @maxedapps

a day ago

@dabit3 Hey, just wanted to say: all the examples you dropped - also in that GitHub repo => just amazing!

0 0 2 898 3

View Details

Maximilian @maxedapps

a day ago

💯

Justin Schroeder @jpschroeder

a day ago

I’ve put a ton of tokens through Jev now, and...I have thoughts. 1. The big labs lost the script TypeSafe’s intro video says, “We’re building prod, not God” — and can I just say? Hell yeah. @typesafeai built something for engineers like me, so I can build products for people

91 119 2K 118K 942

1 1 26 12K 21

View Details

Maximilian @maxedapps

a day ago

@rickasaurus It's genuinely useful. It's that simple

1 0 1 512 0

View Details

Maximilian @maxedapps

2 days ago

@Mng64218162 I think it's simply excitement and playing around with it + trying to push it to the limits. Nothing wrong with that

1 0 0 125 0

View Details

Maximilian @maxedapps

2 days ago

If you want to go viral today on YouTube, create some "anti-Jev" videos. People hate hype, especially related to AI, since there IS way too much noise and hype in that field indeed. Doesn't change the fact that Jev is a big thing, but that won't matter for many (right now, at least).

5 0 28 4K 4

View Details

Maximilian @maxedapps

2 days ago

I think the "Who needs Jev, we had classification models since way before LLMs" takes are amongst the dumbest I've read over the last 24h. Kind of the "who needs Dropbox, you can easily build this yourself" vibes but worse.

16 5 151 8K 11

View Details

Maximilian @maxedapps

2 days ago

Awesome! So many new use cases getting unlocked!

nader dabit @dabit3

2 days ago

Also have been playing with @typesafeai Jev, insane! So many immediate use cases and new apps are possible. What a time to be a builder! Sharing some experiments here starting with: Keystroke oracle / predictive launcher: Your launcher ranks by aliases, fuzzy match, and

96 132 2K 401K 2K

3 7 158 33K 116

View Details

Maximilian @maxedapps

2 days ago

@vavanessadev The good old days! :) Super happy to hear the course was helpful back then!

0 0 1 25 0

View Details

Maximilian @maxedapps

2 days ago

@askadityapandey ?

0 0 0 21 0

View Details

Maximilian @maxedapps

2 days ago

@FrontEndDrake What do you mean?

1 0 0 20 0

View Details

Maximilian @maxedapps

2 days ago

@SevenviewSteve Yeah, video games came to my mind as a great use-case, too!

0 0 5 1K 1

View Details

Maximilian @maxedapps

2 days ago

Jev will unlock many new use-cases and businesses. And we're all in the top 0.001% right now! I think Jev is that kind of release that truly marks another huge evolution step in this (arguably noisy) world of AI. Think of what you can do with Jev + LLMs. Jev + (finetuned) small models. Jev without LLMs at all. We're super early. And imagining that there is so much more to come - with future versions of Jev, competitors. Of course also LLMs, new models. Absolutely mindblowing.

20 23 589 29K 167

View Details

Academind @academind_real

2 days ago

Hi Jev 👋 1. Yesterday TypeSafe AI launched Jev, a model that returns typed decisions + probabilities, not text. 2. LLMs generate tokens for humans. Jev answers predefined questions (Choice / Score / yes-no) in parallel so software can act on the result with no parsing. 3. Fast, cheap, high quality (from what we see now) ~70–500 ms, $0.042 / million input tokens, output free. Early demos and reviews show frontier-level quality on these decision tasks, with vendor speed/cost claims still needing broader independent tests. Only imagining how this affects and speeds up browser and computer use is absolutely insane.

1 3 15 1K 6

View Details

Maximilian @maxedapps

2 days ago

@notkevinzhang Thanks for letting me in :-)

0 0 0 207 0

View Details

Maximilian @maxedapps

2 days ago

@MindVault1 It's truly mindblowing. There's a lot of hype and noise in this industry and Jev isn't any of that as far as I can tell. Now starts the part where we have to figure out how to fit it into the existing tools and best combine it with LLMs => I fully agree with you :)

0 0 2 620 0

View Details

Maximilian @maxedapps

3 days ago

So I'll be honest => My first thoughts about Jev were that it sounds similar to that entire SubQ hype (what happened to that by the way) a couple of months ago. Lots of noise for nothing (?). I genuinely think now, after a few hours of testing, that it's a HUGE missing piece that can really help us build more useful tools and programs when combining it with LLMs. Maybe I'm still wrong and overhyping this but I'm truly hyped about right now, not gonna lie.