David Tompkins
davidntompkins.bsky.social
David Tompkins
@davidntompkins.bsky.social
Developmental psych phd student @ Cornell

Spatial skills | methods | limits to generalizability over time
Incredible. Should yeet apply a not to the condition?
November 7, 2025 at 12:14 AM
That system is also what allows us 8 billion people on the planet. It’s a good deal!
October 20, 2025 at 11:45 AM
Thanks! I’m looking for replications by different authors, so still hard to find! Appreciate the help!
October 18, 2025 at 10:30 PM
The one that gets me is not correcting for each predictor in a linear regression. I never see anyone correct for it, but if you’re willing to report any predictor as significant, you’re inflating the fdr
October 18, 2025 at 10:29 AM
Julian’s for breakfast can’t be beat!
October 3, 2025 at 1:01 PM
And a weird second bias by the way ChatGPT looks like one entity, but CharacterAI looks like many characters. Might be a monogamy vs ai hookup culture effect
September 23, 2025 at 10:44 AM
Amazing news, thanks!
August 21, 2025 at 1:00 PM
Was it much effort to setup? I’ve been ignoring mcp but this is enough praise to tempt me into it
August 21, 2025 at 10:53 AM
Wonder if it works with specific people instead of general concepts? Like could I finetune a model and make a Christmas card with me in one orientation and my partner in another?
August 19, 2025 at 11:38 PM
Right, that does not sound like generative ai? It sounds like grammarly-esque tools. Am I misinterpreting?
July 17, 2025 at 12:47 PM
Surely i did. What did i miss?
July 17, 2025 at 12:40 PM
Surely there are some good uses for generative AI in your organization.

There is no way the optimal amount is 0.
July 17, 2025 at 11:23 AM
Okay but this would actually be fantastic for multiverse analyses
July 17, 2025 at 12:54 AM
Beautiful! Thank you!
July 16, 2025 at 12:38 AM