gk pfp

gk

@gk

306 Following
134 Followers
gk pfp
gk
@gk
🔵 we're excited about provably autonomous agents on /base ai agents are taking on-chain actions, but how do we know if they're really autonomous? with SEQUOAI, we will be showcasing the power of TEEs and on-chain validation to prove genuine ai autonomy.
0 reply
0 recast
0 reaction
gk pfp
gk
@gk
we will be introducing (provably) on-chain autonomous agents with /sequoai soon. hint: trustless execution environments (and on-chain validation)
0 reply
0 recast
0 reaction
gk pfp
gk
@gk
willing to exchange Base mainnet ETH (worth several dollars) for testnet ETH
0 reply
0 recast
0 reaction
gk pfp
gk
@gk
@jessepollak we need 1 Base Sepolia ETH to release the testnet version of our (provably) on-chain autonomous agents. Faucets are not working for us.
1 reply
0 recast
1 reaction
gk pfp
gk
@gk
Was there a specific feature that convinced you to switch completely?
1 reply
0 recast
1 reaction
gk pfp
gk
@gk
"When we launch Private Cloud Compute, we’ll take the extraordinary step of making software images of every production build of PCC publicly available for security research." https://security.apple.com/blog/private-cloud-compute/
0 reply
0 recast
0 reaction
gk pfp
gk
@gk
I’d say AWS Activate (had both Vercel and Stripe credits) and Microsoft for Startups.
0 reply
0 recast
0 reaction
gk pfp
gk
@gk
Short-term: Llama 3 (maybe) and more open-source multi-modal models? More of smaller language models beating the larger language models. Medium-term: Better and more usable AI assistants whether it is wearable or not such as Dot, Tab, Humane Ai Pin. Long-term: AGI? (lol)
0 reply
0 recast
1 reaction
gk pfp
gk
@gk
I love this concept. AI assistants with long-term memory are the new notes/journaling. https://twitter.com/newcomputer/status/1719746992522568010
0 reply
0 recast
0 reaction
gk pfp
gk
@gk
Specialist over generalist. Compact over large models. Cheap over expensive.
0 reply
0 recast
0 reaction
gk pfp
gk
@gk
Very good in terms of context and model size but seems like BAAI/bge-large-en-v1.5 (1024 tokens) is still leading the Massive Text Embedding Benchmark (MTEB) overall for now.
0 reply
0 recast
0 reaction
gk pfp
gk
@gk
Both on-chain excitement (prediction markets) and off-chain excitement (people trying to reproduce it on Twitter) was exciting :)
0 reply
0 recast
0 reaction
gk pfp
gk
@gk
Same happens w/ almost 20. Especially feeling stuck.
0 reply
0 recast
1 reaction
gk pfp
gk
@gk
Lifecycle of my ideas: 1. Find the idea 2. Find a domain 3. Purchase domain 4. Find another idea Result: >25 (unused) domains for now?
1 reply
0 recast
1 reaction
gk pfp
gk
@gk
No problem! I would say Stripe Atlas because of the perks (OpenAI credits, etc.) but there are definitely cheaper alternatives.
0 reply
0 recast
1 reaction
gk pfp
gk
@gk
I heard good stuff about Ramp too, but they do not provide business accounts unlike Brex. Never tried but arc.tech also seems very useful on the banking side, especially with the treasury accounts etc.
1 reply
0 recast
0 reaction
gk pfp
gk
@gk
Banking: Mercury/Brex Invoicing: Stripe Expense management: Brex Credit card: Brex Briefly used Puzzle for bookkeeping, and no experience on taxes side.
1 reply
0 recast
0 reaction
gk pfp
gk
@gk
I recently saw someone doing this :) We've come a long way just in a year
0 reply
0 recast
0 reaction
gk pfp
gk
@gk
Feeling OPtimistic about mainnet
0 reply
0 recast
0 reaction
gk pfp
gk
@gk
Idea: OpenAI GPT-3 + Whisper for cracking any interview. Alternatively, Amazon Transcribe instead of Whisper to have speaker recognition and probably transcribe faster.
2 replies
0 recast
1 reaction