svpino.com
I help companies build Machine Learning • I run http://ml.school. • Posts about what I learn along the way.
370 posts
4,998 followers
158 following
Prolific Poster
comment in response to
post
That’s a pretty good analogy, I think.
comment in response to
post
The training set is meant to train the model. There's nothing nefarious here.
comment in response to
post
These comments are not coming from OpenAI. They are coming from @fchollet.bsky.social who has no incentive whatsoever to hype the model.
In fact, you could argue François' incentive is the complete opposite to keep the ARC benchmark relevant for longer.
Bottom line: Trust him.
comment in response to
post
lol
comment in response to
post
That's my branch, yes.
I use city names to differentiate versions of the product I'm building. The idea of using "Ashville" comes in memory of what happened with the hurricane.
comment in response to
post
In 2021, Codex, then Copilot, then ChatGPT, then Devin, then OpenAI o1...
People keep yelling that "Programming is Dead," and yet the demand for good Software Engineers has never been higher.
Stop listening to midwit people. Learn to build good software, and you'll be okay.
2/2
comment in response to
post
Wow, that’s cool. Do you think the Vision Pro is worth it for that use case?
comment in response to
post
Is this something you prefer using instead of a regular display?
comment in response to
post
I posted about this in the other app, and most replies were about how you'd get "a lot of runtime errors" if you don't work with types.
After 10 years of building enterprise applications exclusively with Python, I'm really confused.
"A lot of runtime errors"? Really?
3/3