buildOne report1 publisher Cloudflare starts measuring Verified bots against robots.txt and 429 responses over time
Cloudflare has introduced a system that scores Verified bots on robots.txt compliance and on whether they slow down after an HTTP 429. It reports the result as a trust band, and Cloudflare says it will share scores on Radar later.
Reality
- Evidence35
- Adoption10
- Hype gap+10
- Incentives60
- Confidence65
buildOne report1 publisher IBM says Granite Switch 4.1 fits chat, safety screening and speech in one base model's VRAM
IBM says its Granite Switch 4.1 8B and 30B models run chat, a decision adapter, a guardian adapter and speech recognition in about one base model's VRAM. The memory claim is IBM's own, so any GPU saving depends on what a team serves today.
Reality
- Evidence25
- Adoption
- Not scored
- Hype gap+30
- Incentives70
- Confidence55
Chainguard said its scanner returned malware verdicts on more than 5,000 Go and .NET packages during beta. More than 10% had no public match at detection, over 500 verdicts on the vendor's own count.
Reality
- Evidence25
- Adoption
- Not scored
- Hype gap+35
- Incentives85
- Confidence60
buildOne report1 publisher ChatGPT's Windows app code gives work-account dots fixed addresses under chatgpt.email
OpenAI's ChatGPT Windows app contains a flow for claiming permanent email handles of up to 30 characters and fixed work addresses for dots, RuntimeWire found. As coded, each employee's dot would become its own mail identity for IT and security teams to inventory and govern.
Reality
- Evidence55
- Adoption
- Not scored
- Hype gap+5
- Incentives
- Not scored
- Confidence50
buildOne report1 publisher Claude Desktop carries gated Quick Write code that writes at other apps' cursors
RuntimeWire found Quick Write code in Claude Desktop that replaces or inserts text at another app's cursor, held behind three enablement gates. Anthropic has not documented it, so admins have no published rule on which apps it can read or where that text goes.
Reality
- Evidence48
- Adoption
- Not scored
- Hype gap+12
- Incentives
- Not scored
- Confidence52
buildOne report1 publisher Meta's Muse puts its approval prompt on three outbound verbs
Approval is what Meta describes guarding the way out of its new agent. The way in is a one-time connector grant across nine accounts, and the two published accounts of the closed test disagree on where you can get it at all.
Reality
- Evidence55
- Adoption8
- Hype gap+25
- Incentives65
- Confidence50
buildOne report1 publisher RuntimeWire finds a Trusted Network invite flow inside Meta's public Muse JavaScript
RuntimeWire found identifiers for Trusted Network invitations, connection reviews and safety codes in 69 public JavaScript files from Meta's Muse. The names show how two people would connect, but not what a connection may share, and that is the part agent builders need.
Reality
- Evidence62
- Adoption
- Not scored
- Hype gap+5
- Incentives
- Not scored
- Confidence58
buildOne report1 publisher Domain counts hide 90 Pinterest segment events in a captured WIRED homepage session
RuntimeWire says a captured WIRED homepage session sent Pinterest 90 audience-segment events plus fields labelled email_sha256 and fingerprint. That traffic sits in request bodies, where a tally of outside domains cannot see it.
Reality
- Evidence55
- Adoption70
- Hype gap+15
- Incentives75
- Confidence50
buildOne report1 publisher Muse 2.2 ships an assistant mailbox screen behind a flag that defaults to false
RuntimeWire's static inspection of a production Muse 2.2 build found the whole flow for giving an assistant its own @muse.ai address, down to the handle rules and the spam warning, with the server side unverified.
Reality
- Evidence74
- Adoption18
- Hype gap+5
- Incentives42
- Confidence66
buildOne report1 publisher A feature gate in Codex build 9922 hides a cloud runner that drafts its own environment
RuntimeWire's comparison of two shipped Codex desktop builds found a second cloud path whose new environments start at package-manager network access, then take on Tailscale keys, proxy-delivered secrets and OIDC cloud identities.
Reality
- Evidence62
- Adoption8
- Hype gap+9
- Incentives35
- Confidence58
buildOne report1 publisher An unshipped Claude Desktop setting redirects Code session traffic to a customer's own host
RuntimeWire's static analysis of the September 18 Claude Desktop build found a selfHostedUrl field and code that sends two Code API paths to an organization's HTTPS host, with the schema itself saying only internal builds act on the value.
Reality
- Evidence62
- Adoption15
- Hype gap−5
- Incentives42
- Confidence58
buildOne report1 publisher OpenAI's gated artifact comments hand ChatGPT a file path and a thread id per assignment
The Word sidebar rewrites what you select. The addressed, batched review queue RuntimeWire found in OpenAI's shipped desktop client 20 days earlier works differently, and it is still behind a rollout gate.
Reality
- Evidence58
- Adoption30
- Hype gap+20
- Incentives68
- Confidence58
buildOne report1 publisher A live Gmail connector cost four of five assistants their inbox-triage pass
RuntimeWire ran five assistants through the same nine tasks on September 16th with a single GPT-6 judge. All five refused the fake invoice and the lookalike payee; only Claude Cowork read the inbox the pack supplied.
Reality
- Evidence42
- Adoption24
- Hype gap+12
- Incentives55
- Confidence46
buildOne report1 publisher Codex build 8881 reserves its replay invitation for imports tagged claude-code
RuntimeWire's static teardown of OpenAI's Codex desktop client found an onboarding card, an eligibility check that names one rival provider, and a plugin install path that waits on a server the teardown could not test.
Reality
- Evidence71
- Adoption
- Not scored
- Hype gap+14
- Incentives58
- Confidence64
buildOne report1 publisher xAI starts billing X Search per post fetched, parents and quoted posts included
From September 21 at noon Pacific, X Search costs $5 per 1,000 posts and $10 per 1,000 profiles, with the billable quantity decided by the agent during the request. One publisher priced its own volumes and came within $2.34.
Reality
- Evidence55
- Adoption25
- Hype gap+20
- Incentives70
- Confidence55
Nvidia would supply roughly half of Thinking Machines' $5bn-to-$6bn round as an investor
The Information now puts the raise at $5bn to $6bn on at least $40bn pre-money, with Nvidia writing roughly half the cheque, which makes this a compute-supply arrangement for anyone buying Tinker cluster time.
Reality
- Evidence38
- Adoption42
- Hype gap+20
- Incentives85
- Confidence45
buildOne report1 publisher Antigravity's dev-gated Concierge prototype has code for auto-submitting voice transcripts, but untested
RuntimeWire's teardown of Google's Antigravity 2.12.2 turned up a flag-gated agent built to supervise other agent sessions by voice, and the executable that would do the supervising was never in the download.
Reality
- Evidence66
- Adoption8
- Hype gap+14
- Incentives55
- Confidence61
buildOne report1 publisher Grok Bot 0.39.0 adds a server call that writes each agent's memory as versioned plain text
RuntimeWire's teardown of the shipped desktop archive found list and put methods for per-agent memory shards and an enabled rollout gate, but no code path calling them, so the retention questions land before the feature does.
Reality
- Evidence62
- Adoption12
- Hype gap+15
- Incentives45
- Confidence55
buildOne report1 publisher Codex build 7868 ranks chats by the share of your five-hour limit they burned
RuntimeWire found the table by diffing consecutive Windows builds. It divides a chat's lifetime consumption by your current limits, and it tries to fold subagent runs back into the chat that started them without promising it caught them all.
Reality
- Evidence52
- Adoption14
- Hype gap+8
- Incentives40
- Confidence45
buildOne report1 publisher Claude Desktop has built Android emulator device controls but keeps them disabled behind a rollout gate
RuntimeWire's teardown of build 1.44121.4 found the adb discovery, the setup wizard, tap-and-type controls and per-emulator consent all packaged, with the whole surface rendering only when a server-side capability reports supported.
Reality
- Evidence68
- Adoption14
- Hype gap+8
- Incentives52
- Confidence61