Joel Martin
@kanaka.bsky.social
Creator of mal/make-a-lisp, noVNC, websockify, clj-protocol, conlink, instacheck, miniMAL, wam, wac, warpy, raft.js (github.com/kanaka).
Interested in Clojure, WebAssembly/wasm, AI/LLMs, JS, Rust, network protocols, web browsers, testing, etc.
Interested in Clojure, WebAssembly/wasm, AI/LLMs, JS, Rust, network protocols, web browsers, testing, etc.
Pinned
Joel Martin
@kanaka.bsky.social
· Nov 18
GitHub - kanaka/mal: mal - Make a Lisp
mal - Make a Lisp. Contribute to kanaka/mal development by creating an account on GitHub.
github.com
If you are into programming languages, or your work/play is polyglot, you might enjoy the mal/make-a-lisp project: github.com/kanaka/mal It is one application (interpreter for the mal Lisp language) implemented in 88 different languages by 36 unique authors.
Free idea for an unsaturated LLM benchmark with large real-world value:
- input: raster image of a diagram
- output: a dot/mermaid diagram with correct nodes, edges, labels, fonts, colors, subgraphs/groupings, etc.
Bonus: it is automatable and has variable complexity.
- input: raster image of a diagram
- output: a dot/mermaid diagram with correct nodes, edges, labels, fonts, colors, subgraphs/groupings, etc.
Bonus: it is automatable and has variable complexity.
December 1, 2025 at 7:24 PM
Free idea for an unsaturated LLM benchmark with large real-world value:
- input: raster image of a diagram
- output: a dot/mermaid diagram with correct nodes, edges, labels, fonts, colors, subgraphs/groupings, etc.
Bonus: it is automatable and has variable complexity.
- input: raster image of a diagram
- output: a dot/mermaid diagram with correct nodes, edges, labels, fonts, colors, subgraphs/groupings, etc.
Bonus: it is automatable and has variable complexity.
Reposted by Joel Martin
Folks: Use AI in hard mode.
Use it to be stronger, not as a substitute for your strength. And for goodness sake, don't let it be your voice. It's a great audience and editor to give you feedback and to train and refine your voice. Your voice is you — don't lose it!
Use it to be stronger, not as a substitute for your strength. And for goodness sake, don't let it be your voice. It's a great audience and editor to give you feedback and to train and refine your voice. Your voice is you — don't lose it!
November 24, 2025 at 10:25 PM
Folks: Use AI in hard mode.
Use it to be stronger, not as a substitute for your strength. And for goodness sake, don't let it be your voice. It's a great audience and editor to give you feedback and to train and refine your voice. Your voice is you — don't lose it!
Use it to be stronger, not as a substitute for your strength. And for goodness sake, don't let it be your voice. It's a great audience and editor to give you feedback and to train and refine your voice. Your voice is you — don't lose it!
Reposted by Joel Martin
Prototypes are the most powerfully tool I know of for increasing the productivity of conversations about what software should do and how it should work
They're tools for thinking
They're tools for thinking
September 21, 2025 at 8:37 PM
Prototypes are the most powerfully tool I know of for increasing the productivity of conversations about what software should do and how it should work
They're tools for thinking
They're tools for thinking
Reposted by Joel Martin
I added a basic tool use chapter to the rlhfbook if you've long heard about it and never spent the time to learn the fundamentals of what it actually is and how the data is formatted.
(I'm becoming more tool-pilled)
(I'm becoming more tool-pilled)
June 16, 2025 at 2:01 PM
I added a basic tool use chapter to the rlhfbook if you've long heard about it and never spent the time to learn the fundamentals of what it actually is and how the data is formatted.
(I'm becoming more tool-pilled)
(I'm becoming more tool-pilled)
Reposted by Joel Martin
I know a lot of my followers are pretty AI skeptical. Which is fine. But for people trying to find a middle way in-between ridiculous waves of pro-AI and anti-AI propaganda you may find this relevant: open.substack.com/pub/mikecaul...
The Apple "Reasoning Collapse" Paper Is Even Dumber Than You Think
We're this far into reasoners and neither hypesters nor skeptics really understand their significance. Also: Read Toulmin.
open.substack.com
June 13, 2025 at 8:51 PM
I know a lot of my followers are pretty AI skeptical. Which is fine. But for people trying to find a middle way in-between ridiculous waves of pro-AI and anti-AI propaganda you may find this relevant: open.substack.com/pub/mikecaul...
Reposted by Joel Martin
"Design Patterns for Securing LLM Agents against Prompt Injections" is an excellent new paper that provides six design patterns to help protect LLM tool-using systems (call them "agents" if you like) against prompt injection attacks
Here are my notes on the paper simonwillison.net/2025/Jun/13/...
Here are my notes on the paper simonwillison.net/2025/Jun/13/...
Design Patterns for Securing LLM Agents against Prompt Injections
This a new paper by 11 authors from organizations including IBM, Invariant Labs, ETH Zurich, Google and Microsoft is an excellent addition to the literature on prompt injection and LLM …
simonwillison.net
June 13, 2025 at 1:35 PM
"Design Patterns for Securing LLM Agents against Prompt Injections" is an excellent new paper that provides six design patterns to help protect LLM tool-using systems (call them "agents" if you like) against prompt injection attacks
Here are my notes on the paper simonwillison.net/2025/Jun/13/...
Here are my notes on the paper simonwillison.net/2025/Jun/13/...
Reposted by Joel Martin
By surveying workers and AI experts, this paper gets at a key issue: there is both overlap and large mismatches between what workers want AI to do & what AI is likely to do.
AI is going to change work. It is critical that we take an active role in shaping how it plays out. arxiv.org/pdf/2506.06576
AI is going to change work. It is critical that we take an active role in shaping how it plays out. arxiv.org/pdf/2506.06576
June 13, 2025 at 2:37 PM
By surveying workers and AI experts, this paper gets at a key issue: there is both overlap and large mismatches between what workers want AI to do & what AI is likely to do.
AI is going to change work. It is critical that we take an active role in shaping how it plays out. arxiv.org/pdf/2506.06576
AI is going to change work. It is critical that we take an active role in shaping how it plays out. arxiv.org/pdf/2506.06576
Reposted by Joel Martin
Lots of neat stuff in this paper showing 30% of US python commits use AI
As of the end of 2024: “the annual value of AI-assisted coding in the United States at $9.6−14.4 billion, rising to 64−96 billion if we assume higher estimates of productivity effects reported by randomized control trials”
As of the end of 2024: “the annual value of AI-assisted coding in the United States at $9.6−14.4 billion, rising to 64−96 billion if we assume higher estimates of productivity effects reported by randomized control trials”
June 12, 2025 at 5:00 PM
Lots of neat stuff in this paper showing 30% of US python commits use AI
As of the end of 2024: “the annual value of AI-assisted coding in the United States at $9.6−14.4 billion, rising to 64−96 billion if we assume higher estimates of productivity effects reported by randomized control trials”
As of the end of 2024: “the annual value of AI-assisted coding in the United States at $9.6−14.4 billion, rising to 64−96 billion if we assume higher estimates of productivity effects reported by randomized control trials”
I posted a new version of my coding agent tutorial: : kanaka.github.io/blog/llm-agent-in-five-steps This version uses the llm library by @simonwillison.net The process is now only 5 steps and the final agent is less than 80 lines of python. Still uses GitHub Copilot (no API/credit card needed).
kanaka.github.io
June 13, 2025 at 12:01 AM
I posted a new version of my coding agent tutorial: : kanaka.github.io/blog/llm-agent-in-five-steps This version uses the llm library by @simonwillison.net The process is now only 5 steps and the final agent is less than 80 lines of python. Still uses GitHub Copilot (no API/credit card needed).
Reposted by Joel Martin
Digital Equipment Corporation no more
Tech giants come and go
nedbatchelder.com
June 9, 2025 at 2:50 PM
Reposted by Joel Martin
Reposted by Joel Martin
Completely agree - code is such an interesting application because you get a basic form of "fact checking" for free, unlike with prose simonwillison.net/2025/Mar/2/h...
Hallucinations in code are the least dangerous form of LLM mistakes
A surprisingly common complaint I see from developers who have tried using LLMs for code is that they encountered a hallucination—usually the LLM inventing a method or even a full …
simonwillison.net
May 8, 2025 at 4:41 PM
Completely agree - code is such an interesting application because you get a basic form of "fact checking" for free, unlike with prose simonwillison.net/2025/Mar/2/h...
Reposted by Joel Martin
I think the one counterexample is that it's good at writing a lot of code, which is creating more text. My rough rule of thumb is that LLMs are good when it's easy to verify the output as correct, which is quite a lot of things
May 8, 2025 at 4:30 PM
I think the one counterexample is that it's good at writing a lot of code, which is creating more text. My rough rule of thumb is that LLMs are good when it's easy to verify the output as correct, which is quite a lot of things
mal/make-a-lisp just hopped up to 89 different languages with the addition of a Hare implementation (harelang.org). Contributed by github.com/einsiedlersp...
The Hare programming language
harelang.org
May 6, 2025 at 12:02 AM
mal/make-a-lisp just hopped up to 89 different languages with the addition of a Hare implementation (harelang.org). Contributed by github.com/einsiedlersp...
Reposted by Joel Martin
New release of LLM, accompanied by a new plugin - you can now use llm-video-frames to turn a video file into a sequence of JPEGs and feed those into a long-context vision model like GPT-4.1-mini simonwillison.net/2025/May/5/l...
May 5, 2025 at 5:53 PM
New release of LLM, accompanied by a new plugin - you can now use llm-video-frames to turn a video file into a sequence of JPEGs and feed those into a long-context vision model like GPT-4.1-mini simonwillison.net/2025/May/5/l...
Reposted by Joel Martin
In which our hero fine-tunes when GitHub actions are run, and battles mysterious YAML and GitHub oddities along the way: nedbatchelder.com/blog/202505/...
Filtering GitHub actions by changed files
How to limit what GitHub workflows run based on what files have changed.
nedbatchelder.com
May 4, 2025 at 2:06 PM
In which our hero fine-tunes when GitHub actions are run, and battles mysterious YAML and GitHub oddities along the way: nedbatchelder.com/blog/202505/...
@emollick.bsky.social You might be interested in this: bsky.app/profile/kana... I made it to be lower barrier than other agent tutorials I've seen: no API sign-up/credits needed (via free GitHub Copilot) and hopefully it is approachable for less experienced python devs (incremental, diagrams, etc)
A 6 step tutorial on creating an LLM coding agent: kanaka.github.io/blog/litellm....
* Uses LiteLLM
* About 140 lines of python when finished
* Can use the free plan of GitHub Copilot (no API sign-up/credit is needed to work through the steps)
* Step 5 of the agent is used to create Step 6
* Uses LiteLLM
* About 140 lines of python when finished
* Can use the free plan of GitHub Copilot (no API sign-up/credit is needed to work through the steps)
* Step 5 of the agent is used to create Step 6
An LLM Coding Agent in 6 incremental steps and about 140 lines of python
I will show you how to create a working LLM coding agent in 6 incremental steps (and about 140 lines of code). We will use the python LiteLLM library and use Github Copilot which means all you need is...
kanaka.github.io
May 3, 2025 at 3:43 PM
@emollick.bsky.social You might be interested in this: bsky.app/profile/kana... I made it to be lower barrier than other agent tutorials I've seen: no API sign-up/credits needed (via free GitHub Copilot) and hopefully it is approachable for less experienced python devs (incremental, diagrams, etc)
Reposted by Joel Martin
The postmortem OpenAI just shared on their ChatGPT sycophancy behavioral bug - a change they had to roll back - is fascinating! My notes here: simonwillison.net/2025/May/2/w...
Expanding on what we missed with sycophancy
I criticized OpenAI's [initial post](https://openai.com/index/sycophancy-in-gpt-4o/) about their recent ChatGPT sycophancy rollback as being "[relatively thin](https://simonwillison.net/2025/Apr/30/sy...
simonwillison.net
May 2, 2025 at 5:04 PM
The postmortem OpenAI just shared on their ChatGPT sycophancy behavioral bug - a change they had to roll back - is fascinating! My notes here: simonwillison.net/2025/May/2/w...
A 6 step tutorial on creating an LLM coding agent: kanaka.github.io/blog/litellm....
* Uses LiteLLM
* About 140 lines of python when finished
* Can use the free plan of GitHub Copilot (no API sign-up/credit is needed to work through the steps)
* Step 5 of the agent is used to create Step 6
* Uses LiteLLM
* About 140 lines of python when finished
* Can use the free plan of GitHub Copilot (no API sign-up/credit is needed to work through the steps)
* Step 5 of the agent is used to create Step 6
An LLM Coding Agent in 6 incremental steps and about 140 lines of python
I will show you how to create a working LLM coding agent in 6 incremental steps (and about 140 lines of code). We will use the python LiteLLM library and use Github Copilot which means all you need is...
kanaka.github.io
May 2, 2025 at 6:20 PM
A 6 step tutorial on creating an LLM coding agent: kanaka.github.io/blog/litellm....
* Uses LiteLLM
* About 140 lines of python when finished
* Can use the free plan of GitHub Copilot (no API sign-up/credit is needed to work through the steps)
* Step 5 of the agent is used to create Step 6
* Uses LiteLLM
* About 140 lines of python when finished
* Can use the free plan of GitHub Copilot (no API sign-up/credit is needed to work through the steps)
* Step 5 of the agent is used to create Step 6
Reposted by Joel Martin
I think the key bit here is that LLMs don't make me X% faster at everything, it makes me able to solve a specific subset of tasks instantly, *including* many tasks that then boost my regular programming. So even though programming is high-leverage, creative, nonfungible, LLMs boost my productivity
folks are talking about how AI makes developers 30% more productive as if that's meaningful—but it simply isn't?
programming is a high-leverage, creative pursuit
it just isn't a matter of outputting fungible programming-units X% faster
programming is a high-leverage, creative pursuit
it just isn't a matter of outputting fungible programming-units X% faster
April 30, 2025 at 7:18 PM
I think the key bit here is that LLMs don't make me X% faster at everything, it makes me able to solve a specific subset of tasks instantly, *including* many tasks that then boost my regular programming. So even though programming is high-leverage, creative, nonfungible, LLMs boost my productivity
Reposted by Joel Martin
Have you sailed "all the seven seas"? Well, what are the seven seas?
April 30, 2025 at 7:03 PM
Have you sailed "all the seven seas"? Well, what are the seven seas?
Reposted by Joel Martin
🔮 The jagged reality of AI – o3’s real world test https://www.exponentialview.co/p/the-jagged-reality-of-ai-o3 (comparing the best models on real-world tasks) #AI
April 24, 2025 at 6:46 AM
🔮 The jagged reality of AI – o3’s real world test https://www.exponentialview.co/p/the-jagged-reality-of-ai-o3 (comparing the best models on real-world tasks) #AI
@simonwillison.net I forked your llm-fragments-github plugin: github.com/kanaka/llm-f... Supports local files/dirs and git working copies. Also adds file path prefix (useful context when working with lots of files).
Question: Can fragment plugins add/use options?
Question: Can fragment plugins add/use options?
April 24, 2025 at 1:36 AM
@simonwillison.net I forked your llm-fragments-github plugin: github.com/kanaka/llm-f... Supports local files/dirs and git working copies. Also adds file path prefix (useful context when working with lots of files).
Question: Can fragment plugins add/use options?
Question: Can fragment plugins add/use options?
The Github Copilot premium requests multipliers (docs.github.com/en/copilot/m...) is probably the most honest information about the real cost of various models. That 50x multiplier on GPT-4.5 is certainly eye watering.
April 22, 2025 at 11:13 PM
The Github Copilot premium requests multipliers (docs.github.com/en/copilot/m...) is probably the most honest information about the real cost of various models. That 50x multiplier on GPT-4.5 is certainly eye watering.
Reposted by Joel Martin
Doing Real Work With LLMs: How to Manage Context https://www.jonstokes.com/p/doing-real-work-with-llms-how-to (this is so good) #AI #prompting #ContextWindows #search
April 19, 2025 at 2:53 PM
Doing Real Work With LLMs: How to Manage Context https://www.jonstokes.com/p/doing-real-work-with-llms-how-to (this is so good) #AI #prompting #ContextWindows #search