bcjordan

10 ms·

61.
by bcjordan 3y ago
Author's writeup about the project: https://medium.com/@iamarunbrahma/fine-tuning-of-falcon-7b-l...
62.
by bcjordan 3y ago
Folks who are here and interested in parallelization for CI/CD may also be interested in Dagger.io — I had heard about it on HN over the years but not played w it. It's basically a more fine-grained Docker-like executor with bette
63.
by bcjordan 3y ago
OK, this is awesome. As a fellow game developer who learned to navigate the complexities of matchmaking and server orchestration the hard way (and begrudgingly), seeing Rivet open source a matchmaker and server orchestrator is exciting and
64.
by bcjordan 3y ago
What a wonderful idea, would love to have this be a monthly thread. Homebrew clubs style—bring the half-baked experiments you've been tinkering with, we'll all learn something from it
65.
by bcjordan 3y ago
I'd be curious to check it out! I've been wondering this last year about the use case of compressing highly repetitive logs in a streaming fashion, and whether fine-tuning a model or combination of LLM / datastore might make
66.
Papers with Code State-of-the-Art Benchmarks(paperswithcode.com)
4 points by bcjordan 3y ago|0 comments
67.
Code Generation on HumanEval Leaderboard(paperswithcode.com)
2 points by bcjordan 3y ago|0 comments
68.
OpenTau – Using Large Language Models for Gradual Type Inference(github.com)
2 points by bcjordan 3y ago|0 comments
69.
by bcjordan 3y ago
> With many men the passing of these extraneous tests is a very grievous interference indeed I feel seen 120 years ago I am hopeful massive LLMs/other models will be an enabler in recognizing and fostering everyone's talents, c
70.
by bcjordan 3y ago
I remember periodically coming across services/platforms that purport to make setting up secure AWS accounts / infra configuration easier and default secure — anyone know what I may be thinking of?
71.
Mentat - The AI Coding Assistant(github.com)
1 points by bcjordan 3y ago|0 comments
72.
DaedalOS(dustinbrett.com)
6 points by bcjordan 3y ago|3 comments
73.
by bcjordan 3y ago
Played with writing a little script to format/sort that: Microsoft takes pains to obscure role in 0-days that caused email breach - https://news.ycombinator.com/item?id=36740133 - (64 comments) Microsoft still unsur
74.
by bcjordan 3y ago
Are there new system requirements known for these?
75.
LangChain – Some quick, high level thoughts on improvements/changes(github.com)
1 points by bcjordan 3y ago|0 comments
76.
by bcjordan 3y ago
Curious if this sort of trained compression could radically reduce the size of log files / analytics storage (or if there exist simple methods for that which perform well already?)
77.
by bcjordan 3y ago
Paper: https://arxiv.org/abs/2306.04050 > We provide new estimates of an asymptotic upper bound on the entropy of English using the large language model LLaMA-7B as a predictor for the next token given a window of p
78.
LLMZip: Lossless Text Compression Using Large Language Models(github.com)
7 points by bcjordan 3y ago|4 comments
79.
Communal computing for 21st century science, Apr 2023 [video](youtube.com)
1 points by bcjordan 3y ago|0 comments
80.
MAS.862: The Physics of Information Technology(fab.cba.mit.edu)
99 points by bcjordan 3y ago|6 comments
81.
by bcjordan 3y ago
Typo in title s/blocky/Blockly
82.
by bcjordan 3y ago
Right, maybe Airtable is to Slack as Sortabase is to Discord
83.
by bcjordan 4y ago
Super cool! I'm curious, does Metal's version support do anything to solve the problem of "I originally embedded with model A, but now I'd like to take my same data and re-embed with a new model B"? I've heard
84.
by bcjordan 4y ago
It sounds a lot like a cache issue (CDN caching something from the API layer? new cache sharding? run out of keyspace?). In addition to users reporting other users' summaries, this user and two others report seeing someone else's
85.
by bcjordan 4y ago
It'd be neat if they had an optional "donate to archive" option which would let textfiles.com attempt to recover any non-personal software/resources from the past (as he wrote in 2011 it's nearly too late to recover
86.
by bcjordan 4y ago
Not sure about their policy/internals myself, but since GPT-3.5-turbo is 1/10th the cost of the previous model it's really hard now to run up a significant (to a funded startup) bill via legitimate usage — especially since at
87.
by bcjordan 4y ago
Now someone's gotta mash up their production dashboard and Apple Health data to map their heart rate spike to receiving their production outage alert
88.
by bcjordan 4y ago
I'm not sure the actual ChatGPT chat interface is capable of measuring perplexity/burstiness yet (unless OpenAI integrated their recent detection work into the actual ChatGPT model which I don't think is the case). Per GPTZer
89.
by bcjordan 4y ago
I'd be curious to hear how the ingestion/search pipeline works — the searches are incredibly fast!
90.
by bcjordan 4y ago
@dang title/link feels a bit misleading now, worth changing to the Google Cache/archive or something?

More ›