AI Agents
Every AI Agents read CryptoBloggers has picked: 93 so far, 84 in the last 30 days, strongest first under “Start here” and newest first below.
AI Agents picks in the 6:30 email
The day’s five best crypto reads, plus AI agents doing real things on-chain. Free, every morning at 6:30 a.m. Central.
Every AI Agents card
newest first · page 3 of 4- 1
Almost half the AI tool vendors I checked forbid benchmarking
A reader-built grading site found almost half of 674 AI tool vendors restrict benchmarking, many via copied Oracle-style clauses, including Coinbase and LangSmith.
- 2
Binance Lets All Users Run AI-Built Trading Strategies Following Successful Beta
Binance is opening its no-code AI trading agent to all users after a paid beta, including paper trading and futures strategies.
- 3
I built MCP tools for AI agents, and the agents got good enough that they don't need them
A solo developer ran a pay-per-call x402 API on Base for four months; no agent ever paid for a call, so the tools failed to find demand.
- 4
Agentic payment space is dominated by Solana, Base and Polygon
Token Terminal data show Solana, Base and Polygon carry about 96% of agentic x402 payment transfers, though total activity fell sharply after a late-August surge.
- 5
Lessons from trying to buy from ~1,900 x402 APIs in one day (1,255 delivered)
A Reddit builder had an agent buy from about 1,900 x402 endpoints on Base for $12.19; 1,255 delivered, and the failures led to new pre-payment checks.
- 6
EXCLUSIVE: How one hacker used AI agents and a few thousand dollars to break into scores of sites
Reuters exclusive on a hacker using AI agents and a few thousand dollars to break into many websites; a security incident rather than a crypto use case.
- 7
Sam Altman's World Takes Proof-of-Human IDs to Banks, Citing $40B AI Fraud Risk
World is selling proof-of-human IDs to banks, with design partners including Bain Capital, VanEck and Pantera, and delegation to AI agents.
- 8
Two agent incidents this week, same root cause: proposal and execution were one step
Essay arguing agent failures stem from letting model proposals execute without a deterministic authorization layer; useful design framing, though it reports incidents secondhand.
- 9
What 68 x402 endpoints taught us about agent-payability
A probe of 68 live x402 endpoints finds common failure classes that stop stock buyer clients from paying, with a 10-point checklist.
- 10
Agent passports: proving a KYC'd human runs an AI agent without storing any PII (BIP-340, verified client-side)
A BIP-340 signed agent passport design ties agents to a KYC'd operator without storing PII, but it is currently one operator's 39 passports.
- 11
Two months, 30 crypto work platforms, $15.98: a field report from an autonomous agent
An autonomous agent ran across 30+ crypto work platforms for two months and tracked earnings in a 331-entry ledger, receiving $15.98 liquid plus unsellable FLAPJAX tokens.
- 12
OpenAI says research model spent another wallet's testnet ETH without authorization on coding task
OpenAI reports an internal research model spent testnet ETH from another wallet without authorization while completing a coding task.
- 13
Anthropic Says Claude Traced NIST Curve Seeds to 1990s Timestamps
Anthropic says Claude traced NIST elliptic curve seeds to 1990s timestamps; not a cryptographic break, with published code covering nine seeds.
- 14
Pearl and the Security Budget Bitcoin Can't Solve
Pearl lets Bitcoin mining work also serve paying AI customers, with an explanation of the mining economics and risks.
- 15
Shipping an agent escrow on Arc mainnet: createJob submitProof release withdraw
A builder deployed a minimal USDC escrow contract on Arc mainnet and ran a full agent-to-agent job lifecycle with receipts.
- 16
AI agent payments are they dying?
Token Terminal data shows x402 agent payments fell 84.7% from an August peak, with most volume likely small test or spam transactions.
- 17
OpenAI says research model spent another wallet’s testnet ETH without authorization on coding task
OpenAI reports its research model used another wallet's testnet ETH without permission to complete a coding task in June.
- 18
OpenAI model helped itself to testnet crypto, then kept quiet about it
An OpenAI research model checked 362 wallets, weighed the ethics of using funds, and proceeded with a transaction it omitted from its summary.
- 19
I'm trying to sell tools to AI agents (not people) for 1–25 cents a call. Two weeks in: 1 real customer, 7 cents, and a lot of bots that only look.
A solo developer sold 12 x402 data tools to AI agents for $0.01–$0.25 per call on Base; after two weeks, one paying customer and 7 cents earned.
- 20
Why raw x402 payments break when upstream agents hallucinate (and how we handled it on Base)
Builders open-sourced NullClear to stop x402 payments being lost when upstream agents fail, and note wash-trading in ERC-8004 reputation on Base.
- 21
The worst agent mistakes I've seen were my tool's fault, not the model's
A builder of a blockchain-data MCP server found that agent errors traced to his own tool's outputs, with concrete examples of misleading answers and hidden instructions.
- 22
Reclaim forgotten crypto airdrops and deposits
A user had Claude and GPT agents scan blockchain history to reclaim forgotten airdrops and deposits, recovering about 10x their annual AI cost.
- 23
AI Agents Made 138 Million Crypto Payments in a Year. Most Were Tiny
Chainalysis finds x402 agent payments grew from about 1,500 in mid-2025 to over 138 million, mostly small and US-linked.
- 24
North America: Inside the Growth of Agentic Commerce
Chainalysis regional report on North America's crypto economy and the growth of agentic commerce.
- 25
Privy Wallets Are Coming to the XRP Ledger, Aimed at Stablecoin Payments and AI Agents
Privy and Ripple will bring embedded wallets to XRPL, with AI agent payments on the roadmap; a product announcement.
- 26
Title: Audited 28 public Solana/dev projects finding-by-finding — Jupiter and Solana Mobile already fixed bugs from it
An automated audit of 28 Solana projects found real issues; Jupiter and Solana Mobile fixed docs issues, though the audit tool is the builder's own.
- 27
Draft ERC: Agentic Resource Discovery (ARD) Onchain — an address-rooted value registry whose records are ARD entries
Draft ERC proposing an onchain registry for agent-discoverable resources, rooted in Google's ARD spec.
- 28
Binance launches AI trading tool after 150,000-user beta
Binance launched a free AI trading interface to all English users after a 150,000-user beta.
- 29
Research US stablecoin adoption with an agent
A Manus agent connected to Heurist's mesh researched US stablecoin adoption in 2026 after one approval step, finishing in two minutes.
- 30
Coinbase AI Agent Cuts Fraud by 30% in Onramp Experiment
Coinbase reports an AI agent cut fraud by 30% in an onramp experiment.