@SuiNetwork@EmanAbio An agent can only quote a business it can trust.
That needs a system designed for agents: the business signs what it says, it's stored on neutral ground, and every agent reads the same source.
That's what we're building on Sui and Walrus.
@kimblgn@satyanadella Agreed. The same holds for what a business says about itself to an agent. It should be a statement the business signed, stored on neutral ground, that any agent can read. That's what we're building on Walrus.
Decentralized AI is usually about compute and open models. @kimblgn of @WalrusProtocol says memory is the next layer. The context you build with AI shouldn’t be locked inside whichever lab’s model you used. It should move with you, on infrastructure no single company controls. @SuiNetwork@Mysten_Labs@liqquidity
good news on catching this early and preventing a disaster
if enough key shares would have been extracted every single multichain asset held by near users (including all zec wrapped on solana by near) would have been instantly drained
(btw projects like @TrezuApp keep insisting they are non-custodial but the scenario above would have instantly drained all of their users’ assets, because the MPC network has full custody, not their users)
this is exactly why we developed 2PC-MPC
vulnerabilities are not going to go away, custodial MPC networks are a honeypot and attract more attackers the more successful they become, they only have to get it right once and all assets of all users are instantly gone
zero-trust MPC is the only way, especially for custody use cases
🇺🇸 Anthropic paid $1.63 MILLION to two Trump-connected lobbying firms, and their ties to Rubio are now drawing attention after the memo that helped bring down fmr. NVIDIA CEO Jensen Huang.
One firm, Continental Strategy, received $740,000 and is led in Washington by Rubio’s former Senate chief of staff. Another partner is Susie Wiles’ daughter.
The other, Ballard Partners, received $890,000 and previously employed Wiles herself.
Federal filings show the firms lobbied offices including the National Security Council and State Department, both overseen by Rubio.
And here's where things get INTERESTING.
Huang was pushing open AI models that could compete with Anthropic’s closed systems. His departure potentially benefits its rivals.
But the filings contain NO evidence that Anthropic orchestrated his removal or that its lobbyists influenced Rubio’s memo.
Source: Zero Hedge / Writer: Sol
Before an agent runs an experiment, agree on what each result means.
@Theoryvc partner Andy Triedman's example: if the new classifier wins 95% of the time, ship it. Under 50%, close the ticket. Anything in between comes to a person.
Full @chain_ofthought episode: https://t.co/VgDafsNyJu
@suidevelopers What's the best pattern you've seen for proving a normal web domain (https://t.co/8zgyhqwuIe) and a Sui address belong to the same owner? SuiNS covers .sui names, not DNS. A DNS TXT record, a well-known file, or something on-chain? Anything people already check against?
@suiwire Would be a great deal if Sui validated the code, the contracts and there was an issurance fund, that made payouts to developers on Sui when things like stoppages occurred, just an idea 💭
DeepSeek open sourced their entire app - APIs for its full suite.
This is the most thorough OSS move. A silent yet loud defiance to centralised closed AI.
A revenge for Jensen and global OSS movement.
What can AI do for you? @vercel CTO @cramforce encourage you to vibe check: paste your text and your rules into a chatbot and see if the model is anywhere up to the task.
If it isn't, wait a couple months for the next generation of models.