Content pfp
Content
@
https://warpcast.com/~/channel/aichannel
0 reply
0 recast
0 reaction
shoni.eth pfp
shoni.eth
@alexpaden
who is discovering the hidden elephant red flags of why nobody in AI is joining? crypto being a scam is already known. Nobody cares about the open data though. why?
4 replies
0 recast
15 reactions
rubinovitz pfp
rubinovitz
@rubinovitz
Why would they join? Theres no research alpha or status or $ to be gained for them here.
1 reply
0 recast
0 reaction
shoni.eth pfp
shoni.eth
@alexpaden
gm8 left because i blocked him because we were both bored doing it to each other, and because we were the only two actually promoting anything that even remotely resembles the pursuit of general intelligence. you think he wants to read 10-20 research papers a week so someone can come and completely overshadow him in his own niche by reposting a template image of "the perfect prompt" guideline? NO. Take that shit to linkedin. Fix the damn niches or your only AI research poster left will be @m-j-r.eth > this isn't directed at you @rubinovitz https://warpcast.com/gm8xx8/0x2f544570
1 reply
0 recast
2 reactions
𒂭_𒂭 pfp
𒂭_𒂭
@m-j-r.eth
tbf I'm a lazy AI research poster used to cast along with gm8 at /ai , but was turned off by the highlander moderator vibes. the most direct way we can incubate upstream discussion is by involving Nous and Prime Intellect among others, upgrading crypto-compatible services like openrouter. most importantly, SOTA is educational, and casts like mine are inadequate for the productive experience that will achieve runaway network effect. we really need instructors like Mollick, Karpathy, Carmack, Jim Fan, et al to discover the possibility of crowdfunding wrappers here that can inspire novelty back into their main gigs. tbqh, I'm fine with watching the industry play out from my shadowy corner, will try my best to continue the async conversation. but it honestly isn't enough.
1 reply
0 recast
2 reactions
shoni.eth pfp
shoni.eth
@alexpaden
well i want(ed) to make a business from novel ai research and despite best efforts i'm unable to do it alone / without a community of likeminded people. re those companies: https://warpcast.com/alexpaden/0x182193f6 i agree about instructors, that's the perfect phrasing too. i'll have to put more effort towards it / that was the original reason i switched channels, to dumb the content down.
2 replies
0 recast
1 reaction
shoni.eth pfp
shoni.eth
@alexpaden
so what can farcaster and merkle do? any thoughts on that? i think it might take a bit of manipulation in the same way autofollow list was manipulation.
1 reply
0 recast
0 reaction
𒂭_𒂭 pfp
𒂭_𒂭
@m-j-r.eth
I wonder about this, not sure what they can do besides focus on improving features as they are already. there could be undiscovered cross-fertilization, like community activities like farwalks could sell VLMs in a "bereal" miniapp. there could be more frictionless geoguessr-esque information about local news. are things like amps fully tuned to sentiment and interests? are the feeds as good as crowdsourced alternatives? idk, there's definitely more app-level, social-level unlocks, but need to ponder that more. besides, we still have to see how 7702 subagents work out, how this supports Ethereum value accrual, how agents self-preserve onchain, how they embody through powerusers, etc.
1 reply
0 recast
0 reaction
shoni.eth pfp
shoni.eth
@alexpaden
these features all seem silly to me, to reference the paper itself “attention is all you need” and these attentions economics are all f*cked
1 reply
0 recast
0 reaction
𒂭_𒂭 pfp
𒂭_𒂭
@m-j-r.eth
hmm, what's your end goal? I want some kind of capability unlock that brute datacenter scale won't accomplish. agreed that attention is cooked, but it's still upstream of novelty.
1 reply
0 recast
0 reaction
shoni.eth pfp
shoni.eth
@alexpaden
I’m not sure I understand the question or your desire, can you specify the context of goal + elaborate on your desire Just curious
1 reply
0 recast
0 reaction
𒂭_𒂭 pfp
𒂭_𒂭
@m-j-r.eth
ultimately, intelligence can be much cheaper, and I expect a public market apparatus will never run out of demand. instead of some opaque, overpriced SaaS that's dubiously capable and some group attractiveness effect, I would love to see the civilizational end-goal where the crypto side figures out optimal signal-gathering incentive -> buying power, while the decentralized AI side serves the cheapest-but-successful predictive control & *motivating* experience that can generalize to all online human circumstances. but as you say, the economics in the short-term is not automatically sustainable, so I'd like to find the novelties that at least unlock some other nodes of the tech tree, even if not completely scalable at first. what do you enjoy on Farcaster (or more generally)?
1 reply
0 recast
0 reaction
shoni.eth pfp
shoni.eth
@alexpaden
Is blockchain or Postgres cheaper? I never really thought about it from wanting it cheaper when decentralized. What do I want from farcaster… to learn new things which act as a flywheel to building on farcaster (identity centric ai)— when I first joined I just wanted people to know who I was and what I liked to theorize about
2 replies
0 recast
0 reaction
shoni.eth pfp
shoni.eth
@alexpaden
Oh interesting I misanswered your question.. What do I enjoy… sometimes it excites me to meet people from different walks of life (musicians actors models)… I’d enjoy learning new things but I haven’t learned much in a while My fault too
1 reply
0 recast
1 reaction
𒂭_𒂭 pfp
𒂭_𒂭
@m-j-r.eth
what's cheaper is whatever's offered with certain conditions, and decentralized drives that relentlessly where postgres will pause in whatever moat. in the longterm, cheap-to-validate environment with optimized, proven lookup will suffice, variable costs at the physical layer might be subsidized out. I say cheaper because it's a big inflection point when "Flippy" is OOM cheaper than the entry-level cook. I expect more humans to experience what today is pricing them out. more experience, more uplifting motivation, more humans perturbing whatever low-confidence models remain. I sympathize with wanting people to know me through interests and theory. tbqh, idk if that's permanent, but these onchain social networks are the best option society gets when the rate of probabilistic control exceeds any institutional capacity. the flywheel could come from anywhere, chances are good it will show itself where the censors can't contravene. idk, just hope everything works out.
0 reply
0 recast
0 reaction