Skip to content

Hermes

Named in the reference to Part 3 ('A Multi-Agent Setup with Hermes') as the multi-agent setup preceding this coordinator search; no further detail is given in the supplied material.

Known aliases

  • Hermès
  • Hermes AI
  • Nous Research Hermes

Current stories

buildOne report1 publisher

Strix, Cairn and Hermes took 600,000 card records from two retailers, Gambit Security says

Gambit Security says an operator chained three open-source AI tools to steal more than 600,000 unexpired card records from two companies. The operator's own review put the average completed scan at $25.46.

Publishers:dev.to

Reality

Evidence55
Adoption20
Hype gap+10
Incentives60
Confidence50
investConfirmed8 publishers

Bloomberg ties LVMH's China weakness to a Louis Vuitton suit against Molly Tea

LVMH's fashion and leather sales fell 5% organically in the third quarter, 3.2 points worse than analysts expected. Bloomberg ties China weakness to backlash over a Louis Vuitton suit against Molly Tea.

Perspective Coverage

8 publishers
Builder
Builder 2%
Operator
Operator 32%
Investor
Investor 66%

Reality

Evidence80
Adoption
Not scored
Hype gap+15
Incentives
Not scored
Confidence78
leadershipOne report1 publisher

The AI trade reaches Korean department stores about a year late

Korean luxury spending has held through a volatile quarter. Bank of America's work says the equity wealth channel reaches store sales roughly a year later, and chipmaker payrolls arrive on a timeline of their own.

Reality

Evidence55
Adoption60
Hype gap+20
Incentives70
Confidence50
productOne report1 publisher

Rabbit's new agent reaches your local files from Telegram and iMessage

Rabbit has stopped making the R1 and moved its agent into a cloud service that touches desktop files through a local node. Jesse Lyu says the handheld sold more than 100,000 units before production ended.

Publishers:wired.com

Reality

Evidence38
Adoption24
Hype gap+34
Incentives78
Confidence44
productOne report1 publisher

Putnam follows Mirror's solo screen with a 24-inch table built for the whole room

Brynn Putnam's new company makes a touchscreen table built for several people at once. The evidence offered that households want one is a WHO loneliness classification and a hotel-sponsored poll about experiences.

Publishers:techcrunch.com

Reality

Evidence44
Adoption
Not scored
Hype gap+36
Incentives66
Confidence56
scienceOne report1 publisher

UHP turns the choice of agent harness into a field in the request

The Unified Harness Protocol specifies how an application starts a task on an agent runtime, follows it, cancels it and collects the files, borrowing the shape of OpenAI's Responses API so existing streaming clients need no changes.

Publishers:unifiedharnessprotocol.org

Reality

Evidence32
Adoption
Not scored
Hype gap+35
Incentives62
Confidence52
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.

Publishers:runtimewire.com

Reality

Evidence42
Adoption24
Hype gap+12
Incentives55
Confidence46
securityOne report1 publisher

Unclaimed packages named in 120 llms.txt files drew beacons from Fortune 500 networks

A scan of 6,214 corporate domains found 120 sites whose AI-facing docs point at code packages and domains nobody owns, and registering a few of those names produced installs traced back to Claude, Codex and Hermes agents inside real networks.

Publishers:schneier.com

Reality

Evidence28
Adoption58
Hype gap+27
Incentives68
Confidence41
buildOne report1 publisher

Bridgeless mode removes the global object that legacy React Native modules look up

The JSI path lets Hermes call C++ host objects directly, which is why RCTBridge has to go. The practical consequence is that your upgrade date is set by whichever vendor SDK still reaches native code the old way.

Publishers:dev.to

Reality

Evidence44
Adoption
Not scored
Hype gap+9
Incentives55
Confidence51
buildOne report1 publisher

Harness choice moved token use 83-fold with the model held constant

The spread traces to two numbers you can read off your own logs: the tokens a harness spends before any work starts, and how many turns it takes. Together they predicted total token use with an R-squared of 0.99.

Publishers:thenewstack.io

Reality

Evidence58
Adoption34
Hype gap+12
Incentives42
Confidence55
securityOne report1 publisher

An agent guard that runs on your laptop, and cannot tell you whether anyone keeps it on

HOL Guard pauses risky actions from Claude Code, Cursor and other coding agents locally. Its privacy design also means the vendor has no retention data to show you.

Reality

Evidence42
Adoption30
Hype gap+12
Incentives62
Confidence52
leadershipOne report1 publisher

An unguarded agent got a founder banned from Resy in 48 hours. Recovery took 32 days.

Brian Distelburger told Business Insider his personal agent stack triggered a Resy deactivation without booking a single table. Recovery ran on goodwill, not support tickets.

Reality

Evidence30
Adoption20
Hype gap+15
Incentives60
Confidence38
productOne report1 publisher

The Empty Box Is Now the Authentication Token, and It Trades Separately

More than 160,000 "empty box" listings on eBay have turned packaging into a priced credential. Credentials that trade apart from the goods they certify become a fraud surface.

Publishers:wired.com

Reality

Evidence52
Adoption58
Hype gap+24
Incentives57
Confidence55
buildOne report1 publisher

The conductor is the bottleneck: local agent stacks fail at orchestration, not at the workers

A dev.to build log finds Gemma 4 26B holds deep single-artifact work but loses the plan after one or two hand-offs, while the coordinators that can plan will not fit in memory.

Publishers:dev.to

Reality

Evidence44
Adoption
Not scored
Hype gap−5
Incentives22
Confidence50