dan-g

10 ms·

61.
by dan-g 3y ago
Generative Agents: Interactive Simulacra of Human Behavior[1]. Make sure to check out the recorded demo! [1] https://arxiv.org/abs/2304.03442
62.
by dan-g 3y ago
The post mentions GPT4All benchmarks—maybe that’s where the confusion lies? https://gpt4all.io/
63.
by dan-g 3y ago
If you’re dead-set on buying online, look for a local boutique somewhere and see if they sell online as well. They generally have good return policies, and in my experience boutiques tend to have much higher quality pieces—generally orders
64.
by dan-g 3y ago
Not quite sure why this deserves a downvote, it’s pertinent since it’s also an open-source flashcard system and there are lots of other comments about it, but okay!
65.
by dan-g 3y ago
Since Anki has been coming up quite a bit in the comments, and this is a self-hostable (but not yet spaced repetition) flashcard system, I wanted to note that Anki does let you run your own sync server now: https://docs.ankiweb.n
66.
by dan-g 3y ago
Which is my point, this didn’t add anything substantive to the discussion. From the commenting guidelines: > Comments should get more thoughtful and substantive, not less, as a topic gets more divisive.
67.
by dan-g 3y ago
Did you have any actual points from the article you wanted to discuss, or call out as wrong? Or are you just communicating that you refused to read it because of the first line?
68.
by dan-g 3y ago
Related: Vicki Boykis has a great blog[1], particularly for those trying to get their feet wet in the ML engineering space—I really found "How I Learn Machine Learning"[2] super useful as a template to tackle new ML problems, as s
69.
by dan-g 3y ago
Nope, that is “Celebrity What-Ifs” on the Midjourney Reddit… The source linked from the tweet is https://bjoernkarmann.dk/project/paragraphica , but it seems to be not responding.
70.
by dan-g 3y ago
Original source: https://twitter.com/BjoernKarmann/status/1663496103998750721
71.
by dan-g 3y ago
For anyone interested (and a Bloomberg subscriber), the version of the article published on Bloomberg has some additional photos: https://www.bloomberg.com/news/articles/2023-05-22/bryan-joh...
72.
by dan-g 3y ago
A list of “Considered Harmful” snow clones: https://en.wikipedia.org/wiki/Considered_harmful
73.
by dan-g 3y ago
I see “Securities Fraud” in the title, I know it’s a Matt Levine article before I even click (Matt famously likes to show that everything is securities fraud in his newsletter)
74.
by dan-g 3y ago
Changed to “piece”— not sure what else to call it. Maybe a profile? But to me that connotes more of a biography or something.
75.
by dan-g 3y ago
Full New York Times piece: https://www.nytimes.com/2023/05/01/technology/ai-google-chat...
76.
by dan-g 3y ago
This is great— do you have a writeup anywhere?
77.
by dan-g 3y ago
Maybe not the interesting, but the most recent— I was annoyed that Bloomberg’s email newsletters didn’t have an associated RSS feed, so I wrote a script that uses JMAP to take the most recent emails from them from my inbox and publish to an
78.
by dan-g 3y ago
I didn’t view it that way—they’re reporting on an event that spurred an FAA investigation. I’d hardly call that muckraking.
79.
by dan-g 3y ago
This is the way. A combination of the two would be prototyping in the language you know best (b), and then once you have it working rewriting it in (a). That way you’re only dealing with one novel thing at a time. For me that’s Python (b) a
80.
by dan-g 3y ago
I feel like this is me spinning in circles deciding on “the best language” to use for a side project, never committing, and then inevitably losing momentum. Frameworks and languages are a software engineer’s “kit”, I guess?
81.
by dan-g 3y ago
Related: Matthew Butterick has a very accessible online book called Beautiful Racket on how to create little languages using Racket to solve problems (including some Advent of Code) which I found a really fun read: https://beauti
82.
Building LLM Applications for Production(huyenchip.com)
9 points by dan-g 3y ago|0 comments
83.
by dan-g 4y ago
Andy Matuschak’s How to Write Good Prompts is another good resource about this: https://andymatuschak.org/prompts/
84.
by dan-g 4y ago
[1996]
85.
On OpenAI's Terrible Arguments Against Transparency(cyberdemon.org)
7 points by dan-g 4y ago|0 comments
86.
Using ChatGPT to generate Git commit messages(cyberdemon.org)
7 points by dan-g 4y ago|2 comments
87.
by dan-g 4y ago
Link: https://www.theatlantic.com/science/archive/2023/03/chernoby...
88.
by dan-g 4y ago
DOOM Emacs addresses this issue (and is an incredible, if opinionated, Emacs “platform”): https://github.com/doomemacs/doomemacs > Gotta go fast. Startup and run-time performance are priorities. Doom goes beyond by
89.
by dan-g 4y ago
(2017) A classic story!
90.
by dan-g 4y ago
Similarly, I’ve recently (re-)realized that if I put my computer away and take pen and paper notes during meetings, I miss fewer details and am more engaged with better retention—even if I just go right to the recycling bin and toss the not

More ›