Paul Han retweetledi
Paul Han
2.3K posts

Paul Han
@pauljunsukHan
Transcribing dexterity @HapticaAI @Yale | @Meta | @Livermore_Lab
San Francisco, CA Katılım Kasım 2014
1.2K Takip Edilen803 Takipçiler

does insta360 put a backdoor in?

Paul Han@pauljunsukHan
the usb hotspot is because the macbook wifi needs to connect to the insta360 and so we need another radio
English

accessed insta360 go ultra debug mode ( upper left corner screen)

jackie@lambdacalcc
We do a little cheeky reverse engineering.
English

the usb hotspot is because the macbook wifi needs to connect to the insta360 and so we need another radio

jackie@lambdacalcc
We do a little cheeky reverse engineering.
English
Paul Han retweetledi
Paul Han retweetledi
Paul Han retweetledi

For the past couple weeks, me and @pauljunsukhan have been running an experiment where we record our entire lives (camera, microphone, computer, even phone capture) so we can teach LLMs everything about our lives.
The results are incredible: LLMs can recursively reason across context and give authentic advice because the context comes from physical reality.
English

i think but for stray dogs for the delivery bots
Tejes Srivalsan@tejessrivalsan
after the overwhelming support for EGO-SNAKE, i’m excited to share EGO-BIRD 100,000 hours of pov bird footage to train the next generation of autonomous drones
English

@SYNESTHEIZURE in 3.13 along with the gilectomy. I was writing a 100Mb/s real time data recorder/control unit in python when 3.12 was just out. JIT can have huge benefits esp if doing bit level protocol decoding.
English

@lauriewired @snwy_me wow what a crossover i show lauriewired’s youtube to everyone i know
English

nah that’s a fair point, it’s prob physically possible to package it.
it’s more of an “does that make any sense” architecture-wise. I respect it, but M-chips are more of a scale-up from a mobile SoC angle (firestorm (aka M1 P cores) iirc was just A14 really). whereas AMD/Intel consumer traditionally “scale down” from server architecture as the main driver.
If you imagine a hypothetical modern X-Serve as “a bunch of mac studios in a box” sure maybe. But an “M3 super ultra” with 100+ cores presented as a single chip ehhh i doubt it
(but who knows, maybe this post will age badly in a few years)
English

apple had a server line (XServe) but discontinued it because they had no advantage over other servers. nobody bought them
with Apple Silicon there's a real reason for Apple servers to exist - imagine an M3 Ultra but designed without desktop thermal/power constraints in mind

snwy@snwy_me
apple *has* to have a successor lined up, given how lucrative Apple Silicon is for ML it would be bad business to not have a real Dell GB300 tower competitor, right? they also really need to start getting these in actual racks
English

@Impish_Bunny float pebble mint have to go down to daly city to buy
English


















