At @axiom_xyz we’ve been building RL environments to help AI models do FV at the binary level and accelerate defensive efforts like these. Our environments are based on real code, scalable, and challenge even frontier models. Get in touch if you are interested!
I didn't predict and do not predict defense running ahead of offense automatically. I predict cybersecurity will be endgame defense-favoring but the transition will be hard. In particular, "good guys find vulns in already-released software before the bad guys do" is not a defense-favoring game. Making the game defense-favoring requires deeper shifts in methodology including formal verification, aggressive TCB minimization, etc, and we've only barely started making those transitions.
(This distinction is a big reason why I continue to be friendly toward slowdowns/pauses, and we need >1% of GDP going toward cyber + bio + physical world hardening)
Millions of @deel users can now earn Morpho powered yield privately.
Idle balances are routed via @tempo Zones before depositing in Morpho Vaults curated by @SentoraHQ, making positions and earnings private.
in a weird way the business side of security companies are not unlike michelin restaurants in that talent is highly specialized and scarce, reputation is on the line every day, must invest in R&D. it's super hard to expand without diluting quality. if you're a boutique auditing firm you're basically in a trap between delivering exceptional work and being expected to scale like a saas company (source: i worked at one)
Phylax is bringing circuit breakers on-chain.
The best way to think about this is policies enforced at the block builder/sequencer level. The predominant worry amongst regulators is how do we protect public on-chain market structure against malicious state updates.
I think we are now converging on a good answer. I would implore everyone to study Phylax as I do believe it's important.
If you're building onchain: the best case scenario right now is that you are the most well-defended protocol in a high-risk environment.
If we looked at security from the lens of public health, it ultimately asks how we make that environment less prone to recurring infection and spread.
Clean air and water treatment does not eliminate the need for doctors but it does alot to reduce dependency.
Onchain systems need the same thing; resilience means being able to withstand attack without relying on flawless incident response every time. As tokenization and AI converge, it's obvious this will be invaluable.
this is our thesis at @phylaxsystems: preserve the need for well-designed protocols and incident responders, while strengthening the population against threats. solving this will create a wellspring that enables new primitives and market designs to take root.
will elaborate on this in a longer post about how it would work: policies ought to remain local, but the capacity to enforce them should sit atop shared infra.
If you were just drained to any of the following addresses, please can you reach out to @SEAL_911 asap. Total loss is >$72M+ and increasing.
TK6DWNpNe1w2iJRNFpU8aHdrPTATxvXT6C
TBkcUMYC7CkTK99tkTnaStQVBastfrs9d9
TCGE3xp6YGRKXxDZiLfysgJW3f22KfMNsW
TSDWtuZ2pARUVz4v3PkL2hi3iXPjowAr5a
bc1qjqgwejnp8dc0x2938x9n9954hj97t82unx49dl
0x83aeac166f6832ae3500000a24510a95a052a599
0x033636e45d519bebb7b5c2520ca6ce56fbdb4f7a
0x69c8f401cfc6cd40ac94691d6d7c48e3b7a47841
this is a parallel to security by obscurity, but for companies whose cash flow assumes people won't trouble themselves with hassles/dark UX. i call it profit by attrition, and it's ripe for disruption
one of the most obvious fintech infra products that should exist right now:
an ai chargeback/dispute agent.
ai agents can already browse websites, make phone calls, reason through policies and negotiate with humans.
so give every cardholder a relentless little lawyer for every transaction they make:
> contacts the merchant
> waits through support queues
> argues why the refund is justified
> escalates across chat/email/phone
> keeps going until the issue is resolved
the infra provider could charge per dispute handled or take a % of recovered funds.
every neobank would rush to integrate this as:
it gives them a real point of differentiation, creates a stronger moat around the product, and makes the customer’s life 10x easier.
then boom, get 50 neobanks running their disputes through your infra and Rain/Stripe acquire you for $50M.
“The excess of liberty… seems only to pass into excess of slavery.” Plato, Republic VIII, 564a
decentralized systems that can’t protect user funds become factories that manufacture demand for centralized ones
the most useful products, in the age of abundant agents ready at hand, will be those that increase our confidence when making highly ambiguous or difficult decisions, by serving the right data at the right time
Consensus view after this year's hacks: DeFi needs junior capital to absorb losses
But nobody says how much (+ more importantly) where it is needed
Can today's lending protocols handle that?
🚨New paper: @Morpho and @Aave (surprisingly!) need more junior capital than optimal