Andrew Phillips
boxtopher.bsky.social
Andrew Phillips
@boxtopher.bsky.social
Engineer, nerd, outdoor sports enthusiast, brewer
I'm not good about finding time to read at home, but I do listen to audiobooks on my way to work.
December 21, 2025 at 8:32 AM
Reposted by Andrew Phillips
I’m glad to see this article get shared some here (non-pw here: archive.is/PHbqH)

I’m very proud of my In-laws for being brave enough to tell their story while many victims understandably are scared to. This kind of #asianhate can be hard to see sometimes. Building awareness to protect is key. 1/2
November 18, 2025 at 6:57 PM
As someone who lives in LA, this is a complete lie.
June 9, 2025 at 7:42 PM
Reposted by Andrew Phillips
Found a baby songbird? Here's a newly updated chart!
May 22, 2025 at 3:01 PM
Reposted by Andrew Phillips
The pope contains potassium benzoate
May 8, 2025 at 5:47 PM
Reposted by Andrew Phillips
Video of the international student at Tufts being arrested by "federal authorities" in Massachusetts has been released and it's terrifying.

They're not even uniformed officers. Just secret police thugs in hoodies and masks.

From WCVB: youtu.be/PuFIs7OkzYY
March 26, 2025 at 5:34 PM
Reposted by Andrew Phillips
March 7, 2025 at 1:14 AM
Reposted by Andrew Phillips
So the new AI from Anthropic is really good at programming. In 2 prompts it made me a Sisyphus arcade game.

Ironically, due to some error, the ball keeps rolling around but you can never touch it.
February 24, 2025 at 8:32 PM
Reposted by Andrew Phillips
I say we beat the shit out of 2025 as soon as it shows up, just as a message to any potential future years
December 31, 2024 at 5:17 AM
Reposted by Andrew Phillips
I've decided to eat too much today
November 28, 2024 at 3:56 PM
Reposted by Andrew Phillips
the human body is not intended to go through three trump election cycles
November 1, 2024 at 2:53 AM
Reposted by Andrew Phillips
This is the fundamental paradox of AI: if it's actually helping you, there is no way to know when it is no longer helping you.

Put another way: if you can supervise it effectively enough to catch its mistakes, you probably didn't need it in the first place.
Scaling up and shaping up large language models increased their tendency to provide sensible yet incorrect answers at difficulty levels humans cannot supervise, highlighting the need for a shift in AI design towards reliability, according to a paper in Nature. go.nature.com/4eCAnis 🧪
September 30, 2024 at 2:31 PM
Unj, it's going to be 2023
Almost the new year, excited for 2021
December 31, 2023 at 4:58 AM
Reposted by Andrew Phillips
The Wrong Stuff xkcd.com/2865
December 9, 2023 at 1:42 AM