Content
@
0 reply
0 recast
0 reaction
ted (not lasso)
@ted
prediction: OpenAI announces and releases ChatGPT 4.5 today
5 replies
4 recasts
39 reactions
Streamm.tv
@streamm
We need an FC bot that creates prediction markets based on casts! Is anyone working on this?
2 replies
0 recast
6 reactions
swampnet
@swampnet
@charleslev
1 reply
0 recast
2 reactions
Charleslev.eth
@charleslev
Yes - or an agent who is using a up-to-date FC data flow to inform its action.
1 reply
0 recast
1 reaction
ted (not lasso)
@ted
this is xAI’s value prop over other models so would be dope to have it for Farcaster cc @dylsteck.eth is this what cortex can do?
2 replies
0 recast
2 reactions
dylan
@dylsteck.eth
that's the goal!! i still have some work cut out for me but some examples of things you can ask on @cortex are: - analyze cast: [cast hash or url] - show me some upcoming events (fetches from events.xyz) - what has @ted been casting about lately? - show me bounties opened in the past ten days lmk if there's anything in particular you'd want to see :) https://withcortex.com
2 replies
1 recast
3 reactions
dylan
@dylsteck.eth
also would add that I assume xAI/grok has some sort of training process *in addition* to on the fly context(rag) about the cast, my approach is a lot more rudimentary rn (grok is the example 100%!)
2 replies
1 recast
2 reactions
Charleslev.eth
@charleslev
rag should be enough - Cortex is operated by agents? meaning - they are extracting those data points for you?
1 reply
0 recast
1 reaction
dylan
@dylsteck.eth
for sure, I guess what I mean is I just haven’t gotten there yet lol rn it’s just a multi-step LLM call that calls tools(via an API I wrote that’s heavily cached) — next steps are to improve the base tools and then add RAG(adding RAG should significantly improve the experience)
2 replies
1 recast
2 reactions
