merc @___merc___
cognitive science phd candidate interested in brains, minds, and machines @binghamtonu new york Joined November 2011-
Tweets6K
-
Followers411
-
Following504
-
Likes15K
“those who can make you believe in absurdities can make you commit atrocities” — Voltaire
sure, but the multi-trillion dollar question is: is better data (idealized, synthetic, etc.) better than more data? i.e., can you get more out of your models with less, but cleaner data? people seem to be pretty split on this…
sure, but the multi-trillion dollar question is: is better data (idealized, synthetic, etc.) better than more data? i.e., can you get more out of your models with less, but cleaner data? people seem to be pretty split on this…
any definition of AGI that is only concerned with “intellectual tasks” is unsatisfactory and conveniently narrow. AGI should also encompass intelligences which give rise to social/emotional and especially physical abilities.
Incredible display of zealot's corner
e/accs and safetyists when the model finishes its training run
SWE-bench is probably contaminated for frontier models (gpt-4/claude-3-opus). Given only the name of a pull request in the dataset, Claude-3-opus already knows the correct function to modify.
we're all going to die, all of us, what a circus! that alone should make us love each other but it doesn't. we are terrorized and flattened by trivialities, we are eaten up by nothing. -- Charles Bukowski
the "debunking devin" video is a painful reminder that: 1.) current AI tech isn't that good 2.) companies pushing AI products will engage in deceit to farm hype (see also Gemini launch demo) 3.) the general public victimized by #2, leaving them blind to the reality of #1
the "debunking devin" video is a painful reminder that: 1.) current AI tech isn't that good 2.) companies pushing AI products will engage in deceit to farm hype (see also Gemini launch demo) 3.) the general public victimized by #2, leaving them blind to the reality of #1
bump this at my funeral 🕊️
Programming is thinking + syntax. Maybe 90% thinking and 10% syntax. While you can automate away syntax, you can only ever outsource thinking to someone/something that can think. Attempts to outsource thinking to a syntax generator do not end well.
I have been working on vision+language models (VLMs) for a decade. And every few years, this community re-discovers the same lesson -- that on difficult tasks, VLMs regress to being nearly blind! Visual content provides minor improvement to a VLM over an LLM, even when these…
I have been working on vision+language models (VLMs) for a decade. And every few years, this community re-discovers the same lesson -- that on difficult tasks, VLMs regress to being nearly blind! Visual content provides minor improvement to a VLM over an LLM, even when these… https://t.co/StilR2HbyO
before long, the majority of text available on the internet will be AI generated. that is a sobering thought. what a time to be alive — for better and for worse.
we are, at our core, storytellers. every decision, attitude and action pens the narrative we gift to the world.
real-time, low latency, speech "clean up" would be a game changer for those with issues in speech production. giving the voiceless an opportunity to be heard may be one of the greatest gifts AI ever gives humanity. that's truly an inspiring thought.
real-time, low latency, speech "clean up" would be a game changer for those with issues in speech production. giving the voiceless an opportunity to be heard may be one of the greatest gifts AI ever gives humanity. that's truly an inspiring thought.