Naved Merchant (@navedmer) header
Naved Merchant (@navedmer) profile picture

Sr. SDE. Post tech memes and opinions. The future is local AI

San Francisco
Joined November 2021
Photo by Naved Merchant (@navedmer)Photo by Naved Merchant (@navedmer)Photo by Naved Merchant (@navedmer)Photo by Naved Merchant (@navedmer)Photo by Naved Merchant (@navedmer)Photo by Naved Merchant (@navedmer)Photo by Naved Merchant (@navedmer)Photo by Naved Merchant (@navedmer)Photo by Naved Merchant (@navedmer)Photo by Naved Merchant (@navedmer)
I built a way to access your local AI models from anywhere using WebRTC - no port forwarding, no cloud providers, completely private. Here's how and why 👇
2
2
111
2tb of vram might be unobtainium right now, but in the future will be commonplace on your personal computers, just like 64gb of system ram was unobtainium in the 90s
Releasing the model weights and technical report of Kimi K3. Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window. New model architecture: 2.5x the intelligence per unit of compute, not just more params. Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale. Model weights: huggingface.co/moonshotai/Ki… Tech report: github.com/MoonshotAI/Kimi-K… Tech blog: kimi.com/blog/kimi-k3
21
ZooCode + GLM 5.2 OpenCode Go plan is all you need. GLM is truly a remarkable mode. It just works :)
29
GLM 5.2 is truly an incredible mode, close to Opus 4.6 in my testing. It has the Claude-esqe quality of just getting what you are trying to get done and gets stuff done! Congrats on creating this amazing model ZAI team!
2
52
Naved Merchant retweeted
The big divide in software development in 2026: Do you do human code reviews? If you do, then "loops", “away from keyboard”, and "multiple concurrent agents" are mostly useless. Why? Because these practices just generate a backlog of code to review faster than humans can review it. Human reviews are the bottleneck, not code generation.
89
20
11
315
62,442
Naved Merchant retweeted
(1/3) Building with Kiro 🛠️ a series spotlighting what the community is shipping. Every week we highlight a builder, a project, and the story behind it. First up: BlueKai, a BlueSky client built for the 200M+ people on $20 flip phones that modern social networks forgot. 👇 spr.ly/6018B6UfmO Shout-out to our amazing judges: @ania_kubow, @svpino, @ssennettau, @ErikCH, @rstephensme, @brianjbeach, @navedmer, @Jay_Raval, @JasonTAndersen, @mrsanfran2, @PhillShaffer
1
6
13
4,465
Naved Merchant retweeted
(1/3) Building with Kiro 👻 Every week we’re spotlighting what people are shipping, starting with our Kiroween hackathon winners. Shout-out to our incredible judges: @ania_kubow @svpino @ssennettau @ErikCH @rstephensme @brianjbeach @navedmer @_Jay_Raval_ @JasonTAndersen @mrsanfran2 @PhillShaffer First up 👇 spr.ly/6015B65DG5
3
2
17
4,136
Even with Opus 4.6 with all its intelligence, I can still confidently say its not replacing engineers yet. It still needs the right context to be effective. Its a force multiplier for sure and i think my productivity has doubled or tripled. But its still a tool like IDEs were.
9
Google will miss a huge opportunity if they dont include this image somewhere in android 17
86
Future Work: 1. MyDeviceAI-Web - browser client 2. Image/PDF support for multimodal models 3. llama.cpp slots for better responses + concurrent inference 4. Seamless desktop updates 5. Custom OpenAI-compatible endpoints
1
35
Everything is Free and Open Source!
20