Profile avatar
whodasam.bsky.social
Data Engineer & AI/ML Product leader. PolPhil, CogSci, Science nerd; Left Libertarian incrementalist. "The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts." Bertrand Russell
405 posts 316 followers 1,282 following
Regular Contributor
Active Commenter
Posts 29 Comments 21

A fun logic puzzle

submitted 35 days ago • 0 comments

Carrying around exploding batteries may become a relic.

submitted 35 days ago • 0 comments

smbc-comics.com/comic/propag...

submitted 35 days ago • 63 comments

they should shut down every social media site and force us all to get a life tbh

submitted 39 days ago • 204 comments

watching tiktok go down I feel like i did when I didn’t bother getting to know a friend’s ex

submitted 39 days ago • 20 comments

submitted 39 days ago • 29 comments

I tested 20 models on the Jetson Orin Nano so you don't have to. I ran Ollama (because of course) with a bunch of smaller models. How does it measure up? Check it out and see. www.jeremymorgan.com/blog/tech/nv... #JetsonNano #AI #MachineLearning

submitted 60 days ago • 0 comments

This German meme was just brought to my attention and I feel obligated to reshare it here Frohe Weihnachten

submitted 64 days ago • 99 comments

Quasi-Magic:Marginal effects at the mean, average marginal effects, and #Stata's margins command www.kai-arzheimer.com/me-at-the-ma...

submitted 64 days ago • 0 comments

George Borjas and I dug into our crowdsourced experimental data to show what he suspected: that researchers' findings on a politically relevant topic followed their political ideology. www.nber.org/papers/w33274

submitted 65 days ago • 2 comments

'The empty raincoat is to me, the symbol of our most pressing paradox. If economic progress means that we become anonymous cogs in some great machine, then progress is an empty promise. The challenge must be to show how paradox can be managed.' from 'The Empty Raincoat' Charles Handy

submitted 74 days ago • 0 comments

What's the polio vaccine done for you recently?

submitted 75 days ago • 0 comments

Not sure how the genie goes back in the bottle. But, I'd love to see Big Tech's collective faces if the legal system decided LLMs weren't a legitimate "derivative" use in many ToS and weren't protected by fair use policies.

submitted 75 days ago • 0 comments

Nothing self-serving in a multi-billion dollar for-profit company pushing for its approach to be subsidized by public institutions. Well, nothing new anyway.

submitted 75 days ago • 0 comments

Probably not news for api touchers, but I found a repo by @why.bsky.team that implements a bunch of simple and less simple feed generators, including ones using ML to classify posts. Pretty straightforward. Wonder how the ML models keep up with firehose throughput though github.com/whyrusleepin...

submitted 76 days ago • 4 comments

Buzzfeed justifying my distain... again

submitted 76 days ago • 0 comments

Add another to the list of "just because you can do something, doesn't mean you should"

submitted 76 days ago • 0 comments

Rockefeller's net worth circa 1920 would be roughly equivalent to $25-30B today. Bill Gates net worth when he became the richest person in the world ~1994-95, would be roughy $30B assuming 3% inflation. Today, the richest person is worth over 10X that.

submitted 77 days ago • 0 comments

How come everything is 5 years off, always? Is there a psychological barrier that makes people more amenable to reality warping when it's framed this way? Far enough to seem possible but close enough to invest? Does it make people imagine regulatory and social hurdles will be eliminated by then?

submitted 77 days ago • 0 comments

Or you are wealthy enough to hire a night nanny... Anyway you look at it, not a great recommendation for genetic engineering.

submitted 77 days ago • 0 comments

Glad to see experimental practice moving toward a more nuanced view of consciousness... "Consciousness, is not simply a binary—on or off, conscious or unconscious—but instead something that can encompass a continuum of different states that involve different kinds of brain functioning."

submitted 77 days ago • 0 comments

GraphRAG is a great method to reduce hallucinations in LLMs. (MSFT has a white paper on the subject earlier this year.) Neo4j has been building graphDB solutions for a decade now and is a frontrunner in the space.

submitted 77 days ago • 0 comments

Recent discussion of the use of LLMs to enable researchers to find connections and areas of opportunity to do research.

submitted 77 days ago • 0 comments

This is the kind of Human Augmented Intelligence I think will be the better outcome of LLM work over the midterm (3-5 years).

submitted 77 days ago • 0 comments

This feels like modus operandi for the next 4 years: If it works it is a win. If it doesn't work, it's held up as another example of bureaucratic red tape that needs to be removed through extreme measures. Also a win.

submitted 77 days ago • 0 comments

Predicting behavior in Bi-directional causation

submitted 78 days ago • 0 comments

How can we understand the electoral breakthrough of far-right parties in countries previously considered 'immune'? By looking at the role of the media and understanding how the far right can enjoy a reputational shield. New paper out on @govandopp.bsky.social 📢 www.cambridge.org/core/journal...

submitted 78 days ago • 5 comments

It's rare you get such a clear illustration of whose lives have value in America and whose don't

submitted 78 days ago • 817 comments

This is kind of cool. If you have a long thread, you type everything into one box and it divides it up into individual posts and posts it for you. bluesky-thread-composer.pages.dev

submitted 78 days ago • 647 comments