the fact that his excuse for the Rabbit R1 not having any defining feature is that there are technical restraints is ridiculous, it's like if SpaceX had a problem with their engines so decided to just build rockets without them
65
sebastian retweeted
last year my friends and I fed 15,000 people
this year our goal is to feed 30,000 people around the world
here’s how you can help
1. every retweet = 1 person fed
2. every 10 likes = 1 person fed
3. every $2 donated = 1 person fed – at San Francisco, CA
sebastian retweeted
i'm putting a small team together to build my next venture in my holdCo (@halationco)
✦ double digit revenue share
✦ get to work w/ exited operators
✦ ideal experience in growth & sales
either we get acquired in 7 months, or it becomes a high cashflow company
who's down?
is it just me or does geeksforgeeks.com seem AI generated at this point
1
84
Isn't this the case for all Apple products vs competitors
So, Mark Zuckerberg finally tried the Apple Vision Pro and compared it to the Meta Quest 3
He says:
- Quest is 7x less expensive yet way more comfortable
- Quest is better for gaming, Xbox, YouTube
- "I'm surprised how much better Quest is better for majority of things"
My biggest takeaway:
The future of ads feels like his IG post.
This is real. No marketing talk, just saying it as it is.
Zuck's Apple Vision Pro feels like I'm listening to a tech reviewer not a guy worth $165B.
I feel like I'm having a convo in a living room with my friend Mark. You feel his enthusiasm.
Reminds me of the Kayne Superbowl ad. The guy spent $7M and took a 30s selfie video. Highly effective. He sold $19M of product within 24h of that video.
So my point is, if you're creating an "ad"
More living room, less high budget.
Be real.
1
81
was looking through the latest Y Combinator batch and the amount of LLM copilot startups is insane
1
62
sebastian retweeted
I'M BUYING!
I'm acquiring SaaS products in the 1-50k MRR range.
DM are open -- if you're interested in exiting 📬
I've heard a few people say this but I will say it again, what is wrong with the comment sections on viral tweets?
90% of the comments have nothing to do with the original tweet, they're just some random videos or memes
63
Day 3 of 100 days of ML paper implementations:
decided to use Tiny Imagenet as full ImageNet is a bit large and will take too long to train on
haven't had much time to work on this, but progressed with my AlexNet implementation, the paper is a lot nicer to read than Lenet
1
72
Day 5 of 100 days of ML paper implementations (yesterday was 4, typo):
Still working on Alexnet, trying to get local response normalization to work (I got a bit confused by the notation)
50
Day 3 of 100 days of ML paper implementations:
wasn't able to post yesterday due to weather complications, started today on Alexnet, getting Imagenet downloaded on a cloud GPU provider
1
97
Day 2 of 100 days of ML paper implementations:
Finished Lenet-5, was too lazy to implement the RBF layer so just used softmax
1
75













