lachlan

I'm always excited to meet people who are optimistic about the future, think that we can be so much more than we are now, and have experience with bringing dreams into the physical world. hyperturing.com

@fishPointer let me innnnn

fishcord server is taking applications + we are now allowing SWEs as part of our latest DEI initiative

i think chips with burnt-in LLMs that run at a very low power will probably result in much of the world around us being unneccesarily intelligent. cheaper to throw that chip and some flash with a readme into an automatic door opener than develop firmware for it.

gf got scam called by someone pretending to be doordash support asking for her 2f code yesterday immediately after she put in an order. seems like restaurant or driver could be selling customer information. fascinating if there’s a larger ecosystem of this

i think doing engineering due diligence for VCs would be very fun

what the fuck

@fangpenlin indeed, that is sometimes the case

@hyprturing When everybody thinks it's a bad time, maybe it's a good time.

i think it is most likely a bad trade to start a CS degree at this point

this applies to other engineering disciplines as well to a lesser extent. will be hard to be a new grad EE student in four years, not just because PCB design will be largely automated. every area will be accelerated in various ways and to various degrees.

_even_ if model performance plateaued tomorrow, i think the ladder has effectively been pulled up already

especially when i’m tired, i feel that my consciousness is a blob of sea foam floating on a violent and roiling ocean. in only the faintest of glimpses can i sense that there might be a sea floor

So Eon put out a more detailed blog post, my takeaways: Vision inputs are based on (Lappalainen, 2024) (connectome-constrained model trained on an optic flow task) and are "somewhat 'decorative'" and "do not currently substantially influence our behavioral outputs" Walking uses "slight modifications to existing NeuroMechFly controllers, trained to imitate the walking behavior of the fly" so, imitation learning, not connectome-driven Brain-body mappings "can be somewhat arbitrarily chosen by hand (as is our case)" "Best understood as a research platform and a demonstration platform" "Should not yet be interpreted as a proof that structure alone is sufficient" I think this is cool work ! bringing together several existing projects in a sensible way, but I have to say the announcement was premature and insufficiently detailed Their detailed write-up: eon.systems/updates/embodi…

@kanzure That would be a very surprising and notable result to me.

@hyprturing Well there's a chance that synaptic weights don't matter

unfortunately difficult to determine whether this is notable or not given the writing style and lack of published or pre-print paper

@stefan_salaices touching grass up in seattle

@hyprturing lachlan why were u not at chess mr

Anyone else take a semester off from uni to work an internship? Do other universities usually charge a THOUSAND dollars in fees just so that you can go off and get real work experience?

@stefan_salaices would not recommend!

Things like this push me closer and closer to just outright dropping out

the first human to dig in the dirt with a stick was the first transhuman

i’ll be out of town but the game goes on!

i hate food i wish i could just inductively charge