We’re publishing three papers (this is the first) addressing core problems in Protocol Learning: multi-participant, low bandwidth model parallel training of foundation models. These papers are workshop papers - they don’t solve the problem (the paper below is in the DP setting) - but they show how we’re thinking about it, and also help to underline the difficulty of solving this. They also contain promising results. You can’t train large models with consumer devices connected over the internet until model parallel is solved, which is what I would like to do.
gmeow, we're at @EthereumDenver
> rsvp to hang w/ us feb 28 w/ frens @HF0Residency, @FenwickWest, @CryptoNomads_gm, @masknetwork
> booth 228 to putt putt for a cute cat plushie
> cool talks happening this week
> garfield testnet went live 👀
📣 Come hang with me Privy Rally and BlockTal to kick off your ETH Denver weekend!
Whether you're a developer building a consumer app, an engineer scouting your next opportunity, or a founder looking to grow your headcount, we’d love to see you there!
#ethdenver
Recruiting is weird. Hey, that job you have now... How do you like it? How much money do you make? Why do you hate your boss? Want to go through a 27 step interview process for bunch of strangers who are going to be judging your entire worth and career during a 30 min zoom? LMK
Like 2 months ago had $10 in my Robinhood account that they wouldn’t let me withdraw for some reason so I bought #doge with it for $0.0531 and forgot about it.
If asked in a job interview, "what is your biggest weakness" a perfectly acceptable response is, "I can see yours is asking useful interview questions."
I wish more people understood that most (good) recruiters (internal and external) *do* try to expand the search criteria for roles, only to get pushback from hiring managers.