back
user profile
mark_l_watson
21,400karma·8,624submissions·August 19, 2009
about
I am an author of 20 books and a practitioner specializing in artificial intelligence, deep learning, natural language processing, and the semantic web. I have 55 US patents. I code in Common Lisp, Clojure, Swift, Python, Haskell, Java, and Scheme. My web site is https://markwatson.com
My recent books can be read for free online on my web site or optionally you can pay for them at https://leanpub.com/u/markwatson
Twitter: mark_l_watson and Mastodon: @mark_watson@mastodon.social
recent activity (8,624 total)
comment
While the over-arc of this message is good (avoid packaged and processed food) I personally don’t like the advice that these are not top tier foods: non-GMO organic whole wheat (i.e., not soaked with …
comment
While I appreciated Meta's early open weight models - large benefit for society, I am in agreement with Yann LeCun recently. Two things can be done: judicious use of LLMs, especially energy effic…
comment
This article struck a personal chord with me: I bought a new MacBook a week ago and installed minimal software on it, specifically I did not install VSCode and I don’t miss it. I use Emacs exclusively…
comment
Very cool, indeed. One nit-pick: Terminus requiring a lot of setup work: Terminus is trivial to use with a rented VPS. But, ptn solves a different problem
comment
Thanks!! I will try that.
comment
Exactly right! Start with a $5/month VPS, then as needed upgrade to $10, $30, $60 a month. Then switch to scalable cloud.
comment
Q: While I agree strongly with the philosophy of this article, and twice I have set up static site generators for blogs hanging underneath my top level personal domain markwatson.com, each time I caus…
comment
I use a Creative Commons share alike license and my eBooks can be read free online. Occasionally some readers purchase my eBooks, and I use sales as a signal for what readers enjoy and that is a guide…
comment
Very nice, thank you. The tests directory is good for testing, and I suggest adding an examples directory with a few very short and complete simple examples.
comment
Thanks Simon, great writeup. It has been an amazing year, especially around tooling (search, code analysis, etc.) and surprisingly capable smaller models.
comment
I think the author is better off self publishing, based on my personal experience: I wrote ten tech books for big publishers (McGraw-Hill, J. Wiley, Springer Verlag, etc.) and I was so happy being a p…
comment
I had to switch from Duck Duck Go browser to Safari, then the Book Recommendation app worked for me. For what its worth, the recommendations were good, evidenced by my having already read half the rec…
comment
exactly right
comment
> in 5-10 years, most entry-level laptops will be able to manage a local 30B sized model I suspect most of the excitement and value will be on edge devices. Models sized 1.7B to 30B have improved i…
comment
Yeah, I hate that also. I just subscribed to Yanis Varoufakis‘s official feed and installed a Safari extension to block everything on YT not in my subscriptions. I think this will work out: use the Yo…
comment
Thanks for the suggestion. I just bought the UnTrap creator’s bundle social media product and tested with YouTube, X, Facebook, and LinkedIn. This is a safari extension. My biggest surprise was how pl…
comment
My wife and I pay for YouTube and the prevalence of AI Slop definitely reduces the value we get for our money. YouTube has done a fairly good job over the years of providing useful recommendations but…
comment
While I agree that Python is where most of the implementation action is, one of the great things about building applications with LLMs is that almost all API providers offer a rich REST interface, and…
comment
A lot of good info, thanks. I have just lightly experimented with Python DSPy and I will probably give your DSPy.rb gem a try, or at least read your code.
comment
I really enjoyed reading the code listings in the article. Many years ago I was a Ruby fanatic, even wrote a book on Ruby, but for work requirements I was pulled to Java and Python (and occasionally C…
comment
Same experience: on my old M2 Mac with just 32B of memory both Qwen 3 30B and the new Nemotron models are very useful for coding if I prepare a one-shot prompt with directions and relevant code. I don…
comment
For what it is worth, I do the same thing you do with local models: I have a few scripts that build prompts from my directions and the contents of one or more local source files. I start a local run a…
comment
The open models are sometimes competitive with foundation models. The costs of Z.ai’s monthly plans just increased a bit, but still inexpensive compared to Google/Anthropic/OpenAI. I paid fo…
comment
+1 It seems to of had the web app portion updated a year ago. And as you say, the application itself looks ‘done.’ I have frequently used Common Lisp over the last 40 years, and I hear comments about …
comment
Me too. I am on a mobile device but I put the article link on my TODO list.
comment
I agree with your decision. I would feel better about an open source solution using local models run with Ollama and LM Studio. Also: Some uses of AI don’t make sense after I think in terms like: how …
comment
Good for Texas. State governments often protect us from the federal government. Many laws that we have now were only passed at the federal level when about 2/3 of states previously passed the sam…
comment
I turned on API billing on API Studio in the hope of getting the best possible service. As long as you are not using the Gemini thinking and research APIs for long-running computations, the APIs are v…
comment
I second this: I have spent about five hours this week experimenting with Nemotron 3 nano for both tool use and code analysis: it is excellent! and fast! Relevant to the linked Google blog: I feel lik…
comment
I agree, adding one point: a better model can in effect use fewer tokens if you get a higher percentage of successful one-shots to work. I am a ‘retired gentleman scientist’ so take this with a grain …