Open Models And Local Inference Are Back In The Game!


Open Models And Local Inference Are Back In The Game!

The Liquid Engineer – Issue No. 44

Open Models and Local Inference are back in the game!

For a few months, it seemed closed models would outpace open models. Thanks to Chinese labs, the race is open again! And with the Framework Desktop, the right hardware is near!

The last months were hard for open model enthusiasts like me. There was a clear and distinguishable gap between the big closed models and the open models. Google’s Gemini 2.5 is such a clever architect and thinker. Claude Opus 4 is, for AI standards, an amazing coder that delivers working solutions on the first shot. Together they formed a team, where I could have architecture discussions with Gemini and successful implementation sessions with Opus 4. But they were both closed and only available through generous or free tiers to get everyone addicted. After years of feeling too dependent on American technology, my appetite to add yet another dependency is zero.

Thanks to Chinese labs like Qwen, Moonshot, and Z.ai, it looks like this is about to change for coding and image generation tasks. I’m trying to have a little bit of vacation myself right now, yet I’m curious to try out these new models and compare them to the known closed ones.

If only there was an affordable machine to run these new models. With the Framework Desktop, the ultimate local inference machine is about to ship: Small, quiet, efficient, powerful, and affordable.

Good times! Enjoy the summer, wherever you are!

What I Learned this Week

“Claude Code has considerably changed my relationship to writing and maintaining code at scale. I still write code at the same level of quality, but I feel like I have a new freedom of expression which is hard to fully articulate.” LINK

The Hater’s Guide To The AI Bubble. I agree there’s a bubble, but I wouldn’t bet on it bursting soon. With coding agents, there’s real, monetizable value. The dream can live on that all these massive investments will pay back. Still, I agree with the general feeling of uneasiness around the incentives of venture capital-backed companies. LINK

I love Benedict’s newsletter and also listened to his podcast for a while. I don’t agree with many of his positions, but it makes me think harder about my own. I like listening to tech history, but there’s too little content about it. VIDEO

What to Print this Week

This newsletter started out on 3D printing. If you haven't had any contact with it, you should, it's great! Here's the most interesting and fun projects I saw last week.

Perfect post-vacation day activity, where all you achieved was reading a few pages, eating, and sleeping. Making your own statue, you earned it!

Make My Statue

These cases are getting more sophisticated. If only I needed them! Happily traveling without a case at the moment.

OneBlade Travel Flexicase

If you think you’ve seen everything around Gridfinity, there’s an upcoming trend named Woodfinity… 😂

Woodfinity Grid 5x5

Hi 👋, I'm Stefan!

This is my weekly newsletter about new technology hypes in general and AI in specific. Feel free to forward this mail to people who should read it. If this mail was forwarded to you, please subscribe here.

Stefan Munz, www.stefanmunz.com
Unsubscribe · Preferences

The Liquid Engineer from OnTree.co

Founder of OnTree.co. Helping you own your AI and escape the sticky, overpriced SaaS trap. Join the movement 🐣

Read more from The Liquid Engineer from OnTree.co

DHH is into Home Servers, too The Liquid Engineer – Issue No. 49 Home servers are back and many cloud computing offerings are a complete rip-off: DHH discovered the same seismic changes this year, and he's a genius marketer. David Heinemeier Hansson, or DHH in short, must live in the same social media bubble as I do, our most important topics overlap this year: home servers are on the cusp of becoming a serious alternative to cloud offerings and the cloud is turning into an expensive joke....

The Lethal Trifecta For AI Agents The Liquid Engineer – Issue No. 43 Simon Willison published a post a month ago, which is already one of the most important blog posts of the year. With the rise of AI agents, the problem described will not change. But we’ll see more practical demonstrations of it, leading to massive problems. The gist is this: There’s a lethal trifecta of risk for AI agents: untrusted content, access to private data, and external exposure. Here’s what each part means and why...

Escaping Groundhog Day with Agentic Coding The Liquid Engineer – Issue No. 42 I did a lot of coding and experiments with Claude Code in the past weeks. Once the initial thrill of the speed wears off, frustration kicks in. I felt like I was in the movie Groundhog Day. The excellent actor Bill Murray wakes up every day in a hostel on the exact same day. The world around him repeats and only he remembers yesterday. He’s stuck in an endless loop and has to experience the same day again and again....