bbstats

Don't take a percent change of a percent change.

the biggest issue is NO WHITE SPACE.

"What a ____ does not ____" is painfully Claude-Coded (no pun intended)

this is great for the 'make a webapp for me and my family' thing

Mine immediately did it correctly?

playing around with 3.8 - it generates non-working code. basically unusable.

Bothering me like crazy that "Use instead: .cat.to(dtype) for int → categorical, .cat.physical() for categorical → int." doesn't give the requisite code example!

This is awful

M-o-MoE

Love to watch movies like this sad reviewer

too bad we'll never know how good it is, since they used a radar plot to show its benchmark scores!

What application/UI are you using deep seek flash high on? Still copilot or something else

The logic for which models stay active when you click on a group of them is extremely not working

First thing I searched was Pokemon cards and found items with bids at 50% higher than market value...either shill bidding or folks who are bidding blindly.

I would pay $8 for this

the subheadings are extremely AI IMHO

I mean...this is a huge company. For them to do something even mildly consumer-focused I'll take as a win.

their green screen is good. that has nothing to do with this video.

... did you watch the video?

inference is not usually slow for classical ML (save lightgbm)

Eating enough pie could help with that

these are called phrases

all this because of a single week?

would love a multi-image version of this.

this is so silly. porque no los dos?

hug of death?

did you read the article?

"previously people could cheat, now they can't" ?? how is that bad.

making a model worse is very easy.

Causality is literally impossible to deduce...