A hardware wallet's whole pitch is being harder to hack than a laptop.
@KeystoneWallet says an attacker got in via a third-party website tool and took buyer data plus the key that proves its devices are genuine. 18,835 unsold units get binned.
The vault held. The shop didn't.
Finally, crypto x AI that isn't a token with a chatbot taped to it.
zkAPI uses Ethereum so your AI usage isn't tied to your billing identity. The provider still sees your prompts and IP address, and the project says so.
Honest beats "fully anonymous." cc @ethereumfndn
Crypto Twitter, 2017: banks are dinosaurs.
@swiftcommunity, 2026: 17 banks piloting live cross-border payments with tokenised deposits on its ledger.
The dinosaurs read the whitepaper. Now the fight is over who runs the rails, and it might not be us.
The new @bubble MCP beta lets you bring your own AI agent and build inside Bubble.
Every software company is about to find out which one it is: a place agents get work done, or a login screen they route around.
Every tokenization pitch: trillions of assets onchain.
@The_DTCC's new onboarding guide asks what never makes the slide: who holds the rights, who runs custody, who handles recovery.
A token with nobody responsible when it breaks is just a support ticket with a ticker.
So @nvidia says its modular AI factories, planned for early 2027, could deploy in as little as 14 days.
Fine print: the site needs power, cooling and networking ready first. That fine print is doing more work than most AI agents.
The AI race is now a race for electricity.
Ethereum devs pulled a testnet upgrade four days before it was due to check how a gas change hits contracts on mainnet. No new date yet.
Meanwhile someone's announcing mainnet for a token that's never seen a testnet.
Guess which I'd trust with money. @ethereum @VitalikButerin
Most agent demos: booking you a restaurant you didn't ask for.
@Visa, @billcom and @Billtrust have a live pilot of AI agents paying supplier invoices a buyer already approved.
Nobody's making a launch video about accounts payable. That's where the money is. cc @rlacerte
Crypto spent a decade telling banks to rip out their systems and rebuild onchain.
@circle and @ACI_Worldwide are working to plug USDC settlement into existing payment rails, starting with Colombia.
Turns out the revolution ships as a plugin. cc @jerallaire
Doomerism is a media psyop giving people a free pass to be lazy.
Yes, volatility. As usual. Israeli vote today, US midterms, UK budget, Hormuz noise. There always is. Markets are pricing that 24/7/365...
The new doomerism treating AI crypto-hacking risks as the biggest problem on Earth is new fear mongering. Cryptography has been upgraded before. AI is already speeding medicine, science and maths along.
Long arc is still up: far less extreme poverty, longer lives, better tools. Markets swing. Progress compounds. We'll keep building making tools for regular people (and machines) to make payments better, cheaper, faster.
skip max effort on terra. artificial analysis has its low and medium settings on the cost frontier and max behind it artificialanalysis.ai/articl…
you can't leak code that's already open source on github, so the "fired cmo dumps postiz's source" post reads like a stunt. if it is, postiz's next real news reads as one too
google gets paid whichever way you read this chart
AI is killing SEO traffic is not completely wrong, but definitely not completely right.
I a new analysis of Similarweb data (10k domains, US + France, 3 years), I found Organic Search fell -13.2% in the US and 24.4% in France over the last 2 years.
HOWEVER, not all of that is due to AI (Overviews). Paid Search has grown strongly over the same time horizon, probably due to more real estate in the SERPs and harder differentiation from organic results.
Also, organic traffic losses are more concentrated in industries like publishing, which lost lost 41.5% in the US and 42.5% in France over the last 2 years.
Lastly, Google's updates are aggressive and often reduce rank performance that can be, at least somewhat, recovered.
So, as much as I see AI reducing clicks left and right, we should not throw the baby out with the bath water.
Full analysis in the new Growth Memo.
figma's dev seats go first when designers edit the component library directly and engineers stop paying to inspect mockups
hours a day is the wrong brag for a chores assistant. mayer told the NYT power users spend hours a day with dazzle nytimes.com/2026/10/25/style…
budget changes look free because the cost shows up days later. meta's help page says a big budget jump can send an ad set back into learning facebook.com/business/help/3…
Something I'm paying more attention to lately:
How sensitive Meta accounts are to frequent budget changes.
In one account, repeated daily budget moves above roughly 10–15% were often followed by 2–3 days of weaker performance.
In another large lead-gen account, more gradual weekly scaling worked materially better than aggressive increases.
Not saying this is universal.
But I think too many media buyers still treat budget changes as operationally free.
They aren't.
The bigger and more mature the account, the more careful I'd be about constantly forcing the system to re-adjust.
agent org charts demo better than they run. anthropic now tells teams to fold specialist sub-agents into one loop, because each handoff loses context claude.com/resources/article…
smart call staggering the 5.5 family. the three models went out over about two weeks, so each release added its own step to that line and kept the story going for a month
bad trade for that client. they swapped Airtable's support team for a single developer, and that developer is now their key-person risk










