Sign in

Mat Miller

@matdmiller.bsky.social
823 followers 2.3K following 65 posts

AI, Engineer, Developer, Traveler, Technologist, @fastdotai Fellow

PostsRepliesMedia
Mat Miller @matdmiller.bsky.social · 19/02/2025
👋
060
Mat Miller @matdmiller.bsky.social · 30/01/2025
🤣
070
Mat Miller @matdmiller.bsky.social · 11/01/2025
I really liked Gradio for AI web apps but it felt limiting. Then FastHTML came along and I finally felt I could write real web apps myself.
0100
Mat Miller @matdmiller.bsky.social · 11/01/2025
Thanks! Jupyter was what drove me to set this up originally as well! Then I crashed my system when away and couldn't recover it so I added PiKVM. And just kept adding from there. It's so easy to do and I feel more secure with additional Cloudflare auth in front of the apps than with port forwarding.
170
Mat Miller @matdmiller.bsky.social · 08/01/2025
Cursor with Sonnet 3.5. I sometimes will use O1 if I want to knock out a lot of code in 1 shot that has a common pattern that’s likely well represented in the training data. Doing things that aren’t well represented in the training data still requires a lot of supervision.
060
Mat Miller @matdmiller.bsky.social · 07/01/2025
Unfortunately yes. This is expected. It’s not symmetrical like fiber.
030
Reposted by Mat Miller
Simon Willison @simonwillison.net · 06/01/2025
Made a tiny tool to help turn a Bluesky handle into a DID: tools.simonwillison.net/bluesky-reso...
Resolve Bluesky handle to DID

UI text box containing simonwillison.net and a Resolve button

Below, the output:

did:plc:kft6lu4trxowqmter2b6vg6z
5817
Mat Miller @matdmiller.bsky.social · 06/01/2025
@howard.fm’s following list is good!
130
Mat Miller @matdmiller.bsky.social · 06/01/2025
True! I use ephemeral VM’s through Databricks for data engineering work and I don’t think of them as serverless but I guess they are. When I think of severless I think of AWS Lambda with near instant starts and short runtimes.
030
Reposted by Mat Miller
Simon Willison @simonwillison.net · 04/01/2025
I like Anthropic's "contextual retrieval" suggestion for creating chunks simonwillison.net/2024/Sep/20/...
simonwillison.net
Introducing Contextual Retrieval
Here's an interesting new embedding/RAG technique, described by Anthropic but it should work for any embedding model against any other LLM. One of the big challenges in implementing semantic search …
0162
Mat Miller @matdmiller.bsky.social · 06/01/2025
Welcome! Glad to have you!
010
Mat Miller @matdmiller.bsky.social · 06/01/2025
Yea. Serverless is cool for small tasks but just doesn’t work for full applications without a lot of consideration if it works at all.
100
Mat Miller @matdmiller.bsky.social · 06/01/2025
For sure! This is perfect for a college budget!
000
Reposted by Mat Miller
Simon Willison @simonwillison.net · 05/01/2025
Looks like @graze.social exists already! Their programming language is a nested JSON object of filters and Boolean operations: www.graze.social/docs/editor-... bsky.app/profile/graz...
{
  "filter": {
    "and": [
      {
        "attribute_compare": [
          "embed.external.uri",
          "==",
          "https://www.youtube.com/watch?v=E8Ew6K0W3RY"
        ]
      },
      {
        "regex_matches": ["text", "\\bthe\\b"]
      },
      {
        "regex_negation_matches": ["text", "\\bunwanted_term\\b"]
      },
      {
        "user_network": ["devingaffney.com", "in", "follows"]
      },
      {
        "text_similarity": [
          "text",
          {
            "anchor_text": "This is an important update",
            "model_name": "all-MiniLM-L6-v2"
          },
          ">=",
          0.3
        ]
      },
5704
Mat Miller @matdmiller.bsky.social · 05/01/2025
For starting out and especially for personal projects it just makes a lot of sense. You just don’t have to make an ‘investment decision’ every time you want to try something out. It’s ~free so just follow your curiosity!
110
Mat Miller @matdmiller.bsky.social · 05/01/2025
Quarto is fantastic!!! Let me know if you have any questions 😀
020
Reposted by Mat Miller
Mat Miller @matdmiller.bsky.social · 01/01/2025
🔥 Host your apps for FREE and simplify your infrastructure! 🚀 Tired of complex cloud hosting and the recurring bills that come with it? Host locally on hardware you already own and access anywhere with a browser. It's simpler than you think! Learn how in my latest post. 🔗👇
3345
Mat Miller @matdmiller.bsky.social · 01/01/2025
And no, this isn't sponsored or anything, I just love @Cloudflare Tunnels and appreciate the generous free tier!
000
Mat Miller @matdmiller.bsky.social · 01/01/2025
blog.matdmiller.com/posts/2024-... #selfhosting #cloudflare #homelab #developers @Cloudflare
blog.matdmiller.com
Self-Hosting Your Applications for Free: A Guide to Cloudflare Tunnels – Mat’s Blog
110
Mat Miller @matdmiller.bsky.social · 01/01/2025
Quick start: 1. Point your domain to Cloudflare 2. Spin up the Cloudflare tunnel with Docker 3. Add your apps 4. Access them securely from anywhere Questions about setting this up? Drop them below - I read every comment and love helping others get started!
111
Mat Miller @matdmiller.bsky.social · 01/01/2025
I built an app that has helped my 2-year-old daughter start to read, apps to control 3D model software, and an app to stream local TV stations over the internet, along with a few others.
100
Mat Miller @matdmiller.bsky.social · 01/01/2025
I've built half a dozen apps using FastHTML so far over the last 6 months ranging from a hundred lines to several thousand and it's been a delight to work with.
100
Mat Miller @matdmiller.bsky.social · 01/01/2025
Want to build your own apps to run locally? You need to try FastHTML fastht.ml by @howard.fm. It's super easy and fast to get started. I struggled learning web development in the past, and this framework finally made it simple enough for me to really get started.
fastht.ml
FastHTML
Modern web applications in pure Python
110
Mat Miller @matdmiller.bsky.social · 01/01/2025
The performance is impressive too - we're talking 500+ Mbps speeds, more than enough for most applications. Want the really good news? @cloudflare Tunnels has a generous FREE tier and is super simple to use!
100
Mat Miller @matdmiller.bsky.social · 01/01/2025
I've been running this setup for over a year, and it has transformed how I run apps and services. No more worrying about recurring fees, security exposure, and complex setups.
100
Mat Miller @matdmiller.bsky.social · 01/01/2025
Here's what makes it awesome: • Run everything on your own hardware - zero monthly costs • Access from any web browser - no VPN needed • Enterprise-grade security through Cloudflare • Automatic SSL certificates • Takes just minutes to set up
100
Mat Miller @matdmiller.bsky.social · 01/01/2025
What can you host? Pretty much anything with a web UI: • Your own private ChatGPT instance using Open Web UI • FastHTML Apps • Jupyter notebooks on your GPU rig - like a private Google Colab • Home network monitoring • NAS and file storage • Development environments
100
Mat Miller @matdmiller.bsky.social · 01/01/2025
I've used this for personal projects, but it's also ideal for business use. Many of the same insights still apply! blog.matdmiller.com/posts/2024-...
blog.matdmiller.com
Self-Hosting Your Applications for Free: A Guide to Cloudflare Tunnels – Mat’s Blog
120
Mat Miller @matdmiller.bsky.social · 01/01/2025
🔥 Host your apps for FREE and simplify your infrastructure! 🚀 Tired of complex cloud hosting and the recurring bills that come with it? Host locally on hardware you already own and access anywhere with a browser. It's simpler than you think! Learn how in my latest post. 🔗👇
3345
Mat Miller @matdmiller.bsky.social · 25/12/2024
I’m doing a bit of both! I can’t sit still for this long but I am enjoying spending extra time with the family.
120
Reposted by Mat Miller
Benjamin Warner @benjaminwarner.dev · 22/12/2024
This week we released ModernBERT, the first encoder to reach SOTA on most common benchmarks across language understanding, retrieval, and code, while running twice as fast as DeBERTaV3 on short context and three times faster than NomicBERT & GTE on long context.
27415
Reposted by Mat Miller
Jeremy Howard @howard.fm · 19/12/2024
I'll get straight to the point. We trained 2 new models. Like BERT, but modern. ModernBERT. Not some hypey GenAI thing, but a proper workhorse model, for retrieval, classification, etc. Real practical stuff. It's much faster, more accurate, longer context, and more useful. 🧵
19620147
Reposted by Mat Miller
Simon Willison @simonwillison.net · 19/12/2024
I figured out a prompting pattern for getting Claude to produce fully self-contained Python scripts that execute with "uv run" using PEP 723 inline script dependencies - and now I can one-shot useful Python utilities with it simonwillison.net/2024/Dec/19/...
simonwillison.net
Building Python tools with a one-shot prompt using uv run and Claude Projects
I’ve written a lot about how I’ve been using Claude to build one-shot HTML+JavaScript applications via Claude Artifacts. I recently started using a similar pattern to create one-shot Python utilities,...
820826
Reposted by Mat Miller
Remi Cadene @remicadene.bsky.social · 15/12/2024
HOT 🔥 fastest, most precise, and most capable hand control setup ever... Less than $450 and fully open-source 🤯 by @huggingface, @therobotstudio, @NepYope This tendon-driven technology will disrupt robotics! Retweet to accelerate its democratization 🚀 A thread 🧵
37327
Mat Miller @matdmiller.bsky.social · 14/12/2024
Hey @runme.dev - love your notebook-style script execution! Any plans for Windows native support? Many of us in corporate environments need this without the WSL dependency. As a Jupyter notebooks fan, I can immediately see the value of this but a many of my use cases today would be Windows native.
000
Reposted by Mat Miller
David Bikard @dbikard.bsky.social · 13/12/2024
Excited to share that we have just released genomenotebook 1.0 together with @trichomedoctor.bsky.social ! If you like to analyze genomics data in Jupyter notebooks this is for you. You can get beautiful genome browsers directly in your notebooks: dbikard.github.io/genomenotebo...
dbikard.github.io
genomenotebook
A genome browser in your Jupyter notebook
37722
Reposted by Mat Miller
Jeremy Howard @howard.fm · 13/12/2024
Non-commercial license, as usual from MSR. 🙄 And of course they don't even give you the option to pay for it.
6563
Reposted by Mat Miller
Ian Maurer @imaurer.bsky.social · 08/12/2024
Chat with any open source repo easily. Gitingest (free online tool) turns any GitHub repository into a single markdown file for pasting. Claude artifacts makes this 300k token output pretty easy to work with.
46412
Reposted by Mat Miller
Simon Willison @simonwillison.net · 06/12/2024
This is a neat new variant on RAG - no vectors, not even full-text search, instead showing the model a header hierarchy and giving it a tool to read the relevant sections My notes here: simonwillison.net/2024/Dec/6/r...
simonwillison.net
Roaming RAG – make the model find the answers
Neat new RAG technique (with a snappy name) from John Berryman: > The big idea of Roaming RAG is to craft a simple LLM application so that the LLM assistant …
1216020
Reposted by Mat Miller
Nathan Cooper @ncoop57.bsky.social · 05/12/2024
As R&D staff @ answer.ai, I work a lot on boosting productivity with AI. A common theme that always comes up is the combination of human+AI. This combination proved to be powerful in our new project ShellSage, which is an AI terminal buddy that learns and teaches with you. A 🧵
77416
Reposted by Mat Miller
Jeremy Howard @howard.fm · 05/12/2024
I can't begin to describe how life-changing this new project, ShellSage, has been for me over the last few weeks. ShellSage is an LLM that lives in your terminal. It can see what directory you're in, what commands you've typed, what output you got, & your previous AI Q&A's.🧵
518521
Reposted by Mat Miller
Ben Burtenshaw @benburtenshaw.bsky.social · 03/12/2024
For anyone interested in fine-tuning or aligning LLMs, I’m running this free and open course called smol course. It’s not a big deal, it’s just smol. 🧵>>
932363
Reposted by Mat Miller
Simon Willison @simon.fedi.simonwillison.net.ap.brid.gy · 29/11/2024
Amazing demo by Vaibhav Srivastav of structured data extraction running on an LLM that executes entirely in the browser (Chrome only for the moment since it uses WebGPU). simonwillison.net/2024/Nov/29/struc…
simonwillison.net
Structured Generation w/ SmolLM2 running in browser & WebGPU
Extraordinary demo by Vaibhav Srivastav. Here's Hugging Face's [SmolLM2-1.7B-Instruct](https://huggingface.co/HuggingFaceTB/SmolLM2-1.7B-Instruct) running directly in a web browser (using WebGPU, so requires Chrome [for the moment](https://github.com/gpuweb/gpuweb/wiki/Implementation-Status)) demonstrating structured text extraction, converting a text …
0236
Reposted by Mat Miller
Thomas Capelle @capetorch.bsky.social · 29/11/2024
After seeing all those toxic replies to @danielvanstrien.bsky.social , I created a small script to auto-block people. Let's use AI to moderate the replies. gist.github.com/tcapelle/00a... cc. @eugeneyan.bsky.social
2142
Reposted by Mat Miller
Thomas Capelle @capetorch.bsky.social · 29/11/2024
Hello World from Python SDK
051
Reposted by Mat Miller
Stella Biderman @stellaathena.bsky.social · 28/11/2024
A dataset of 1 million or 2 million Bluesky posts is completely irrelevant to training large language models. The primary usecase for the datasets that people are losing their shit over isn't ChatGPT, it's social science research and developing systems that improve Bluesky.
825139
Reposted by Mat Miller
François Fleuret @francois.fleuret.org · 29/11/2024
I find strange that engineering the gradient radically to learn better is not a whole field.
9535
Reposted by Mat Miller
Simon Willison @simon.fedi.simonwillison.net.ap.brid.gy · 29/11/2024
TIL how to implement GitHub OAuth for an otherwise completely static website (HTML+JavaScript on GitHub Pages) using Cloudflare Workers: til.simonwillison.net/cloudflare/wo…
24410
Reposted by Mat Miller
Nick Tsergas @nicktsergas.ca · 28/11/2024
✨ More Bluesky tools for power-users, curious people, and nerds. ✨ One thing that makes this place a better option for anyone wanting a more conscious social experience is its "opensource-ness." It gives us a level of freedom we haven't had since the MySpace era. I missed that. 🧵 1/10
Screenshot of part of SkyZoo's analytics dashboard (see thread)
113373154
Reposted by Mat Miller
Jeremy Howard @howard.fm · 28/11/2024
Did you know that 99% of email today is spam? Your inbox isn’t 99% spam because AI is used to filter it. The same 99% will happen here too, but if AI researchers continue to get perma-banned for making available the datasets needed to filter it, it’s going to make this platform unusable.
4251064