Peer now supports creator payouts from @BagsPay to CashApp
Creators get paid in dollars. Buyers on Peer get to onramp from Cashapp to USDC at up to a 25% discount
BagsPay now works with @CashApp
Launch a coin and share fees with any cashtag.
Creators get paid directly in dollars.
No wallet required. No unclaimed fees.
alaa retweeted
Did you know that when you deploy a token on Bags (even through J7, Rapid, Proxima, etc.) you earn 25% of the fees generated by that token, regardless of who you share fees with?
Go claim your fees...
We're partnering with @BagsPay to help creators receive fees directly in dollars
With Peer, there is no single point of failure. Every payout is a transaction that is private, permissionless and peer-to-peer
Creators can now cash out to X Money. Soon, Venmo, Revolut and many more payment apps
alaa retweeted
Selling my @SolanaFndn Breakpoint 2026 General Admission ticket.
$490 or best offer — transferable via Luma
DM me
#Solana
Every web3 app should integrate OHTTP into its stack
We’re announcing the closed beta of a self-serve Cloudflare OHTTP Gateway. We’re also renaming our Privacy Gateway to Cloudflare OHTTP Relay to better distinguish the two products. cfl.re/4ARJPKX #BirthdayWeek
3
601
Understanding this fundamentally changes how one should approach learning new stuff, what to prioritize and what to disregard as a waste of time (eg syntax), giving you an edge in the next chapter of tech
AI won’t replace you if it can’t fool you, no matter what frontier model they bring.
In serious systems work, the bottleneck is not implementation generation. its specification.
Fix the semantics, invariants, canonical representation, dependency structure, cost model, and proof obligations.
deeper the system, the harder the semantic specification becomes. engineer who defines the semantics, invariants, representation, and proof obligations defines the space in which AI can search.
Defining a frontier specification for a substrate is way harder than building a frontier model. If frontier models are made from frontier specifications, they can replace entire categories of human work. So, yeah, they aren’t truly frontier until they can write frontier specifications themselves. That’s why specification is the real hard problem.
Also real problem is deriving the minimum sufficient computation from the dependency structure, while preserving observational equivalence.
And composition requires a shared semantic representation. Without one, independently generated abstractions cannot reliably compose or optimize globally.
So the engineer’s advantage is not writing more code than AI . just they need defining the mathematical world in which AI’s code is allowed to be correct.

















