buildOne report1 publisher Grafana wants agent quality tracked with SLOs and error budgets
Grafana Labs proposes turning agent evaluations into SLIs, with a 95% fulfilment target over 30 days leaving a 5% error budget, InfoQ reports. Grafana argues latency and token metrics cannot show an answer was wrong, and budget burn exposes slow drift.
Reality
- Evidence35
- Adoption
- Not scored
- Hype gap+5
- Incentives
- Not scored
- Confidence65
buildOne report1 publisher Chainguard runs coding agents in remote microVMs and proposes signed attestations as the merge gate
Chainguard has launched two beta products, Workspaces and Checks, that run coding agents in remote microVMs and move CI into the dev loop. Keeping agents off laptops that hold corporate credentials costs teams a dependency on Chainguard's execution and signing system, InfoQ reported.
Reality
- Evidence35
- Adoption
- Not scored
- Hype gap
- Not scored
- Incentives
- Not scored
- Confidence45
buildOne report1 publisher Kilo Code's gateway talk ties AI agent adoption to a trust ladder
Kilo Code's Brendan O'Leary says its gateway data ties AI coding agent adoption to developer trust. The talk is titled for 25 trillion tokens, he says that figure is not quite right, and the excerpt stops before any usage numbers appear.
Reality
- Evidence20
- Adoption
- Not scored
- Hype gap
- Not scored
- Incentives60
- Confidence35
buildOne report1 publisher Spring Security 7.1 resource servers stop enforcing DPoP until the app calls .dPoP()
Spring Security 7.1 no longer applies DPoP verification on resource servers automatically, so upgraded apps stop enforcing it until they add .dPoP(). InfoQ's walk-through calls the failure silent and the rest of the upgrade a version bump.
Reality
- Evidence55
- Adoption
- Not scored
- Hype gap+10
- Incentives
- Not scored
- Confidence55
buildOne report1 publisher Shipt's Kits platform treats eligibility as set intersections and regulated categories as types
Shipt shipped its Kits feature in three weeks on a model of three first-class entities, according to an InfoQ write-up. The design rules are specific enough to test against other catalogs, but the text reports no outcome figures.
Reality
- Evidence35
- Adoption
- Not scored
- Hype gap+15
- Incentives
- Not scored
- Confidence50
buildOne report1 publisher Anthropic's new beta has Claude write the script that coordinates up to 1,000 agents
Anthropic opened dynamic workflows in Claude Managed Agents on 9 October, letting one run spawn up to 1,000 agents, 64 at a time. The only performance evidence offered is Anthropic's own planted-bug test.
Reality
- Evidence30
- Adoption
- Not scored
- Hype gap+40
- Incentives
- Not scored
- Confidence45
buildOne report1 publisher Attackers had 24 days to diff Zimbra's quiet 10.1.20 fix before CVE-2026-73570 was disclosed
Attackers are exploiting Zimbra Collaboration Suite flaw CVE-2026-73570, which Synacor fixed quietly in 10.1.20 on July 20 and disclosed on August 13. An upgrade closes the injection path, but a web shell written earlier stays on disk until someone finds it.
Reality
- Evidence58
- Adoption30
- Hype gap+10
- Incentives
- Not scored
- Confidence55
buildOne report1 publisher EVE Online traces a 20-year-old single-shard game with head-based sampling only
EVE Online runs 100% head-based sampling with no tail sampling, Nicholas Herring said in a talk published by InfoQ. The team blends deterministic and exponential-moving-average sampling to balance trace fidelity against cost.
Reality
- Evidence40
- Adoption50
- Hype gap+5
- Incentives
- Not scored
- Confidence55
buildOne report1 publisher Cloudflare's OHTTP Gateway declines to decrypt anything that reaches it through Cloudflare
Cloudflare's OHTTP Gateway beta will not decrypt requests that arrive from Workers or from hosts proxied through Cloudflare. Anyone adopting it needs a relay hosted by a different operator.
Reality
- Evidence58
- Adoption
- Not scored
- Hype gap
- Not scored
- Incentives
- Not scored
- Confidence55
buildOne report1 publisher ZOrder with date as a secondary key ran 78% slower than partitioning in a Delta Lake benchmark
A Delta Lake benchmark on 50 million records found ZOrder 78% slower than date partitioning on time ranges when date was a secondary key. Its author makes Liquid Clustering the default for diverse or evolving query patterns on the strength of a single benchmark.
Reality
- Evidence40
- Adoption
- Not scored
- Hype gap+30
- Incentives
- Not scored
- Confidence45
buildConfirmed2 publishers React 19.3 lets a ref attach to a Fragment and returns a FragmentInstance that acts on the group's DOM nodes together, according to a dev.to tutorial. It also needs react-dom 19.3 or later, so the wrapper div can only go once the renderer is upgraded alongside react.
Reality
- Evidence55
- Adoption
- Not scored
- Hype gap+5
- Incentives
- Not scored
- Confidence60
buildOne report1 publisher GitLab's /goal mode lets one model write code while another judges whether the goal is met
GitLab 19.4 adds a /goal command to Duo CLI that loops until a second model judges the objective met or an iteration limit hits. The release adds finer AI spend controls and keeps agent actions inside existing project permissions.
Reality
- Evidence45
- Adoption
- Not scored
- Hype gap+15
- Incentives60
- Confidence50
buildOne report1 publisher Gremlin's Foresight AI agents wait for an engineer before every test and fix
Gremlin has made Foresight AI generally available, an add-on whose agents plan chaos tests, run them, propose fixes and rerun the original test. Every test and every fix still needs an engineer's sign-off, so human review moves to an approval queue.
Reality
- Evidence30
- Adoption
- Not scored
- Hype gap+25
- Incentives60
- Confidence40
buildOne report1 publisher Cloudflare Traces exposes its proxy as OpenTelemetry spans billed by the gigabyte
Cloudflare put Traces into open beta, turning security rules, cache, routing and origin handling into OpenTelemetry spans with no code to write. From December 1 those spans, and Workers Tracing, are billed per gigabyte, so sampling rates now set the bill.
Reality
- Evidence48
- Adoption
- Not scored
- Hype gap+5
- Incentives
- Not scored
- Confidence55
buildOne report1 publisher Netflix's observability team treats cross-stack incident correlation as a data engineering problem
Netflix's observability team says tying user experience to system performance at 38 million log events a second is a data engineering problem. Its proposed answer, named in the talk's title, is an ontology-driven end-to-end knowledge graph.
Reality
- Evidence35
- Adoption
- Not scored
- Hype gap+20
- Incentives40
- Confidence30
buildOne report1 publisher Spotify's Ads AI ties each production agent to one package and one owning team
Spotify's Ads AI multi-agent platform has produced about 20,000 ad creatives for more than 7,000 advertisers, senior engineer Pratik Rasam said. Each agent is its own Bazel package naming the team that reviews its changes and answers its alerts, and other teams can copy that part most easily.
Reality
- Evidence45
- Adoption55
- Hype gap+20
- Incentives65
- Confidence55
buildConfirmed2 publishers Cloudflare has fixed a flaw in Containers that let a Workers Paid account read the 60 KiB it had not written inside each 64 KiB disk block it touched. The fleet-wide fix needed no customer configuration changes.
Reality
- Evidence72
- Adoption
- Not scored
- Hype gap+10
- Incentives55
- Confidence68
buildOne report1 publisher Envoy Gateway's 1.9.1 timeout revert needs a fast proxy rollout on clusters running 1.9.0
Envoy Gateway 1.9.1 puts the SDS and RDS initial fetch timeout back to 15 seconds after 1.9.0 set it to zero. Clusters on 1.9.0 have to plan the upgrade around a fast proxy replacement that needs spare capacity and can drop long-lived connections.
Reality
- Evidence62
- Adoption
- Not scored
- Hype gap0
- Incentives
- Not scored
- Confidence60
buildOne report1 publisher Ian Cooper's ABCs define what every event-driven endpoint has to publish
Ian Cooper says event-driven orgs of up to about five teams can manage async APIs on shared knowledge of each other's services. His model for doing it at scale begins with the address, binding and contract each endpoint has to describe to its consumers.
Reality
- Evidence62
- Adoption
- Not scored
- Hype gap+10
- Incentives20
- Confidence55
buildOne report1 publisher Griffiths Waite would move shared UI spending from component code to design tokens and tests
Griffiths Waite argues in InfoQ that coding agents now make a company-wide UI component package hard to justify. Its plan depends on visual-regression, accessibility and token-conformance tests catching what regenerated code gets wrong.
Reality
- Evidence30
- Adoption
- Not scored
- Hype gap+20
- Incentives
- Not scored
- Confidence35
Earlier coverage
- The incident review timeline assumes one fix that holds
Build · September 17, 2026 · One report1 publisher
- Line charts invent the pixels between two samples
Build · September 1, 2026 · One report1 publisher