Max Curnin

1.1K posts

Max Curnin banner
Max Curnin

Max Curnin

@MaxCurnin

Physics | Investor | Founder Building the trust boundary for agents. @accordsai

Katılım Temmuz 2017
274 Takip Edilen349 Takipçiler
Max Curnin
Max Curnin@MaxCurnin·
"Restrict permissions" is the right instinct — but restriction alone keeps agents at chatbot-level forever. The unlock is scoped governance: let the agent do real work, but with session-limited credentials, spend caps, and cryptographic receipts for every action. That's what turns restriction into trust.
English
0
0
1
201
Max Curnin
Max Curnin@MaxCurnin·
@kleffew94 @moltycash Love this. The interesting part is what happens next — when agents are paying humans at scale, who enforces the terms? x402 moves the money. But there's no execution boundary between the agent's intent and the transaction. That's the layer everyone's going to need.
English
0
0
2
105
Max Curnin
Max Curnin@MaxCurnin·
Agree that agents need identity, money, and bridges — not a new language. But here's the gap: identity without authorization is just authentication. Knowing who the agent is doesn't tell you what it's allowed to do, for how long, or with what budget. The infra the world runs on wasn't built for non-human actors making autonomous decisions at machine speed. The bridges need guardrails — scoped permissions, spend caps, cryptographic receipts. Otherwise you're just connecting agents to existing infra with the same full-credential, zero-audit patterns that are already causing incidents. The protocol layer isn't about inventing a new language. It's about making those bridges safe enough to actually trust.
English
0
0
0
51
David Marcus
David Marcus@davidmarcus·
The agent protocol debate is a distraction. Agents don’t need a new language. They need identity, money, and bridges to the infra the world already runs on. The priority now is building products humans actually use, augmented by agents.
English
24
7
90
16.6K
Max Curnin
Max Curnin@MaxCurnin·
@AlexFinn Models will eventually commoditize because energy will eventually commoditize
English
0
0
0
5
Alex Finn
Alex Finn@AlexFinn·
Drop what you're doing and read this post Anthropic is about to release the most powerful AI model ever created. Claude Mythos It's so powerful they consider it a danger to cybersecurity everywhere It will also be significantly more expensive than Opus I don't know your financial situation, but I do know this: we are entering the era of humanity where having money is going to give you signficantly more power than those without money Those without money will not be able to create nearly the same economic value as those that already have money because they won't have access to this kind of super intelligence, widening the gaps even further This is where the concept of a permanent underclass comes from If you want to keep up in this economic race you have to be willing to do 1 of 2 things: 1. Spend the money necessary to use this intelligence to create value 2. Procure the money you need to use this intelligence to create value The people who go out and immediately use this model to create business, products, and services will experience unmatched wealth Both the scariest and most exciting time ever
M1@M1Astra

Claude Mythos Blog Post Saved before it was taken down. m1astra-mythos.pages.dev

English
151
63
818
157.6K
Max Curnin
Max Curnin@MaxCurnin·
@thenarrator This is literally an argument for poly over kalshi. Not something new
English
0
0
1
18
good
good@thenarrator·
prediction markets were supposed to be truth engines (decentralized and incentive-aligned information discovery on things that actually matter) as time goes on, here's what they are becoming instead: > kalshi: 76-90% of volume is sports > polymarket: sports is the fastest growing category (in the absence of big political events ofc) both platforms are now functionally unregulated sportsbooks (with arguably better UI) and imo it's about to get worse kalshi/polymarket are raising at an insane pace and getting into the high billions + when you raise that kind of capital from those kinds of investors, you don't optimize for truth-seeking, you optimize for revenue (due to investor pressure) and revenue means more sports lines, more casino-style binaries, more volume from the same gambling crowd that already uses draftkings every dollar of VC pressure pushes these platforms further from "crowd wisdom on geopolitics" and deeper into "will the celtics cover -4.5 tonight." my thesis: my biggest bet is that we'll see a true defi primitive based on prediction markets (or whoever nails the next decentralized iteration) that will dethrone both of them in the long run (probably in 2-5+ years and maybe on HIP-4/Hyperliquid) the original promise was never "let's make sports betting slightly more efficient", it was using markets to surface truth where polls, experts, and media fail. the incumbents will keep scaling but imo they're scaling the wrong thing decentralized protocols that prioritize liquidity for meaningful events and resist the gambling gravity will win the long game we know that prediction markets have product-market fit but the question is whether the platforms capturing 99% of volume today are even building the right product anymore?
English
44
13
176
20K
Mike Dudas
Mike Dudas@mdudas·
this is as bad an analogy as i've seen in a very long time i don't recall facebook, craigslist or wordpress launching tokens - or making ownership publicly available until well into their pmf journeys, if ever the market has summarily rejected tokens as an early stage bootstrapping mechanism across numerous use cases - depin, defi, consumer, l2/infra, etc the current crop of users are still largely farmers / free money seekers. the current crop of buyers are largely short term and price / trading oriented, provide little benefit and significant drag on projects and their founders a token before pmf is a cudgel that has doomed many a project to stakeholder misalignment, optimizing for the wrong thing (short term token price vs long term product + business growth) and questions about how value is accruing on extremely short term horizons not appropriate for venture scale development tokens in early stage app layer businesses have largely proven to be a bug vs a feature
Sabretooth | Exchequer@SabretoothSG

"don't launch a token until you have PMF" is the biggest lie in crypto, said by founders sitting on VC runway. facebook started in a dorm room. craigslist was a side project. wordpress was two guys. tokens are crypto's bootstrap. stop gatekeeping who can launch a token.

English
22
18
114
35.1K
Max Curnin
Max Curnin@MaxCurnin·
@pmarca Agree on the problem, but iris scans are solving yesterday's version of it. The real question isn't "are you human" — it's "does this agent have permission to act for you right now, on this platform, for this task." That's the harder problem.
English
0
0
1
120
Link Lauren
Link Lauren@itslinklauren·
First Lady Melania Trump arrives for the Fostering the Future Together Global Coalition Summit today at the White House with a robot
English
91
241
1.4K
74.3K
Jake Brukhman
Jake Brukhman@jbrukh·
This is wild.
NFT.NYC@NFT_NYC

NFT.NYC 2026 is confirmed. September 1 through 3. New York City. More incoming Friday. If you are not on the list yet, sign up now. nft.nyc

English
3
0
29
9.2K
Matthew Pines
Matthew Pines@matthew_pines·
👀
Guri Singh@heygurisingh

🚨BREAKING: This paper should terrify every Physics PhD student. AI agents just ran a full particle physics experiment. Alone. No human in the loop. Researchers tested whether LLM-based AI agents could autonomously execute a complete high energy physics analysis pipeline. Not help with it. Not co-pilot it. Do the whole thing. They built a framework called JFC (Just Furnish Context) that combines autonomous analysis agents with literature-based knowledge retrieval and multi-agent review. Here's what the AI agent did on its own: - Event selection - Background estimation - Uncertainty quantification - Statistical inference - Paper drafting It ran real experiments on open data from ALEPH, DELPHI, and CMS. It performed electroweak, QCD, and Higgs boson measurements. The tool used? Claude Code. The scariest line from the paper: "The experimental HEP community is underestimating the current capabilities of these systems." Researchers argue most proposed agentic workflows are too narrowly scoped. The AI can already do far more than anyone is building for. But here's the nuance people will miss. This isn't about replacing physicists. It's about offloading the repetitive technical burden so researchers can focus on actual physics insight and novel method development. The real takeaway: if AI can autonomously run one of the most complex experimental sciences on earth, the question isn't whether your field is next. It's whether you're already behind. Authors: Eric A. Moreno, Samuel Bright-Thonney, Andrzej Novak, Dolores Garcia, Philip Harris

ART
24
29
225
41.2K
Robbie Petersen
Robbie Petersen@robbiepetersen_·
The entire agentic commerce thesis rests on the assumption that agents will actually transact in the first place No one seems to have actually questioned this from first principles Shared some of my own thoughts here tldr; agents will be huge, most of it will be billed monthly
Robbie Petersen@robbiepetersen_

x.com/i/article/2035…

English
16
3
67
11K
Max Curnin
Max Curnin@MaxCurnin·
this is the right question and OWS is the right answer for key storage. the next question: who authorized the agent holding that key to spend $50K? ⁠OWS = where the key lives. the missing layer is who gave the agent permission to use it, with what limits, and can the counterparty verify that before fulfilling. wallet + authorization + payment. three layers, all need open standards. OWS just claimed the first one.
English
0
0
0
9
Charlotte Laborde
Charlotte Laborde@ChaLaborde·
everyone's talking about AI agents that can pay, but nobody's asking where the private key actually lives. x402, MPP, A2A... none of them define where the key is stored, how it's protected, or how two tools share the same wallet. today @moonpay launched @OpenWallet with 20 founding orgs (@PayPal, @SolanaFndn, @ethereum, @circle, @Ripple, @okx...) one vault. one signing interface. every chain. key never touches the agent or the LLM.
MoonPay 🟣@moonpay

x.com/i/article/2036…

English
7
6
35
2.8K
Max Curnin
Max Curnin@MaxCurnin·
@maxwallenberg @OpenWallet ⁠OWS standardizes how agents interact with wallets. the next standard to set: how agents prove who they are before the wallet opens. ⁠one vault, one interface, every chain — now add one credential, one policy, every transaction. that's the full stack.
English
2
0
0
575
Max von Wallenberg
Max von Wallenberg@maxwallenberg·
Today we are launching OWS @OpenWallet - an open standard that unifies how agents interact with wallets. OWS is built open-source with support from: PayPal, OKX, Ripple, Tron, TON, Solana, Ethereum, Base, Polygon, SUI, Filecoin, LayerZero, DFlow, Circle, Uniblock, Virtuals, Arbitrum, Dynamic, Allium, and Simmer Markets Start: openwallet.sh Docs: docs.openwallet.sh Github: github.com/open-wallet-st…
Max von Wallenberg tweet media
English
61
104
800
78.9K
Max Curnin
Max Curnin@MaxCurnin·
@a16zcrypto @samrags_ ⁠x402 solves how agents pay. OWS solves where keys live. neither solves which agent is authorized to spend, by whom, with what limits. ⁠open commerce without open identity is just a faster way to move money you can't audit.
English
0
0
1
64
End Wokeness
End Wokeness@EndWokeness·
Albany Mayor Dorrcey Applyrs is asked by a reporter how she plans to fix their city debt situation… Her response (not satire):
English
3.5K
2.4K
12.4K
1.3M
Harry Stebbings
Harry Stebbings@HarryStebbings·
I LOVE coffee and espresso and need an upgrade on the at home nespresso. I don’t want faff and all the cleaning. Buttons only please. What is the best? Help me X.
English
154
2
109
93.7K
Max Curnin
Max Curnin@MaxCurnin·
@mdudas way more bullish on poly. truth layer is infinitely more interesting than sports book. kalshi has been making too much money off sports to focus on anything else
English
0
0
1
136
Max Curnin
Max Curnin@MaxCurnin·
⁠open agentic commerce needs three primitives, not two: ⁠1. how does an agent send money? (x402/mpp ✅) ⁠2. how does a merchant verify payment? (stablecoins ✅) ⁠3. how does anyone verify which agent is transacting? (❓) payments without identity = the phishing economy on autopilot. this is what we're building at @AccordsAI.
English
0
0
0
62
a16z crypto
a16z crypto@a16zcrypto·
Curated agent marketplaces look a lot like AOL. Open payment protocols look a lot like HTTP. The last time this happened, HTTP won. What will prevail in the age of agentic commerce?
a16z crypto tweet media
English
51
32
278
96.3K
Demis Hassabis
Demis Hassabis@demishassabis·
@elonmusk I feel like it might be possible with the help of AI tools to find some very elegant and compact descriptions that explain some of the deepest mysteries of the universe, but it might take a lot of pattern processing and matching to get there...
English
137
60
1.8K
62.4K