gandalfgeek

11 ms·

91.
by gandalfgeek 4y ago
If you remember the Transmeta chips from the early 2000s they literally were compilers. Took in x86 binaries but JITed them to an internal RISC instruction set.
92.
Using LLMs for UI interaction [video](youtube.com)
2 points by gandalfgeek 4y ago|0 comments
93.
Enhancing large language models with vision [video](youtube.com)
3 points by gandalfgeek 4y ago|0 comments
94.
by gandalfgeek 4y ago
On MacOS I've been using homerow.app which does the same thing as Scoot but looks a bit more polished.
95.
by gandalfgeek 4y ago
Just me, or does this read like a Godfather-style warning? "Nice AI product you got over there. Would be a shame if something happened to it." But what's different about this AI product/hype cycle compared to every other
96.
LLMs Are Compilers(vivekhaldar.com)
4 points by gandalfgeek 4y ago|0 comments
97.
by gandalfgeek 4y ago
Just came here to say big thanks to the Homebrew crew. First thing I install on a new Mac. There's so much thankless toil that goes into maintaining package infrastructure that we all just take for granted. Thank you!
98.
A Spiritual Successor to Emacs(vivekhaldar.com)
2 points by gandalfgeek 4y ago|0 comments
99.
by gandalfgeek 4y ago
What they're describing in the paper goes much further than LangChain. Toolformer is integrating API calls into the training and inference of an LLM. Short video explainer here: https://youtu.be/ZCdqfIuT81A
100.
by gandalfgeek 4y ago
If you have the time to watch it, this Munk Debate on "Don't Trust the Mainstream Media" is a good articulation of both sides. https://www.youtube.com/watch?v=3vkgROIINEs At the beginning, the audience was ev
101.
Prompt LLMs for code, not arithmetic(vivekhaldar.com)
1 points by gandalfgeek 4y ago|0 comments
102.
by gandalfgeek 4y ago
The paper is fascinating from a compilers + VM perspective. Short video summary: https://youtu.be/ES7fkec3Jv8
103.
Read a paper: Program-aided language models(youtube.com)
1 points by gandalfgeek 4y ago|0 comments
104.
by gandalfgeek 4y ago
The "COVID deniers", as they were labeled, (think people like Jay Bhattacharya, Scott Atlas et al), warned about all of this. They presented the tradeoffs and probabilities on both sides. But they were completely shut down without
105.
by gandalfgeek 4y ago
I was surprised to see the omission of a prompt technique called program-aided prompting. Paper: https://arxiv.org/abs/2211.10435 GitHub: https://github.com/reasoning-machines/pal tl;dr -- LLMs ar
106.
by gandalfgeek 4y ago
You can use https://github.com/zerolfx/copilot.el to get Copilot completions in Emacs. Video explainer: https://youtu.be/dZMGH_3UdSE
107.
by gandalfgeek 4y ago
LambdaPrompt looks really cool. Love the Pythonic expression as good ol' functions, which makes the code read very naturally. What applications were you envisioning when you built it?
108.
Read a Paper: Prompt Engineering for Solving CS1 Problems(youtube.com)
2 points by gandalfgeek 4y ago|0 comments
109.
by gandalfgeek 4y ago
Founded by the guy who used to to head up Google Ads engineering (Sridhar Ramaswamy). I know some might jump to conclusions just from that, but this EconTalk appearance by Ramaswamy goes into his motivations. He sounded like he had soured o
110.
Using LLMs for complex programming [video](youtube.com)
1 points by gandalfgeek 4y ago|0 comments
111.
by gandalfgeek 4y ago
You win this thread.
112.
Can we automate scientific reviewing? [video](youtube.com)
2 points by gandalfgeek 4y ago|0 comments
113.
The Automatic Corporation(blog.vivekhaldar.com)
2 points by gandalfgeek 4y ago|1 comments
114.
by gandalfgeek 4y ago
"Jugaad" is a cultural attitude that is hard to explain outside the Indian context. (Even literally, as the "d" sound in the word has no counterpart in English phonetics. It's similar to "dh" but not quite
115.
by gandalfgeek 4y ago
Back in the early 2000s, Sun ("the network is the computer") had a similar solution that worked seamlessly for most of their software org-- the Sun Ray. https://en.wikipedia.org/wiki/Sun_Ray It was a network
116.
by gandalfgeek 4y ago
Every service has the risk of introducing a new point of failure, of course. But compare dev productivity lost due to service downtime with that of each new SWE in your org burning time to A) setup their own unique snowflake of an environme
117.
How to evaluate LLM performance [video](youtube.com)
1 points by gandalfgeek 4y ago|0 comments
118.
Read a Paper: Automatic Prompt Engineering(youtube.com)
2 points by gandalfgeek 4y ago|0 comments
119.
How truthful are large language models? [video](youtube.com)
1 points by gandalfgeek 4y ago|0 comments
120.
by gandalfgeek 4y ago
Eng mgr at a FANG, was IC previously. The following is assuming the senior folks (managers or senior ICs) around you are acting in good faith. If you actually have toxic senior people, that's a whole different scenario. Managers have t

More ›