Governed AI. Outcomes assured. Operate. Build. Govern. Assure. Banks · capital markets · life sciences. 21 years.
US, Canada, Europe
- Tweets 1,791
- Following 753
- Followers 672
- Likes 451
Same word. Different object.
Their assurance asks whether the data was retained. The examiner's asks whether the work met the requirement.
One is about the asset. The other is about what happened to it.
itmethods.com/insights/answe…
1
20
The filing is not the work.
CRA Article 14 starts 11 September. Filing is a notice. A shared fix is not proof it landed.
@paulgoldman1 in The Trust Layer.
itmethods.com/insights/the-f…
23
The filing is not the work.
CRA Article 14 starts 11 September. 24-hour early warning. The industry will treat it as a filing problem. That work is real. It is not the exam.
A shared fix is not proof it landed. Can you produce the artefact?
itmethods.com/insights/the-f…
Made with AI
77
Monday: a forty-company alliance, a contract playbook, and a TV appearance. Four weeks after the partnership that set it up.
Palantir and NVIDIA are right about where control lives. Ownership of a layer is a deal. Proof of an action is an artefact.
Full piece in The Trust Layer: trustlayer.itmethods.com/p/o…
61
0 reverts. 922 autonomous merges.
Our own production. The number is not the point.
The point is we can show how it was produced.
- Which agent claimed the ticket.
- Which tests passed.
- Which human approved the merge, at what time, under which policy.
A revert count is a claim. An evidence ledger is proof.
itmethods.com/dark-factory
1
159
Three claims your AI vendors are selling. What each actually proves:
Air-gapped → isolation. Where the work happened, not whether it was allowed.
Never trained on → a policy, until it is a contract term with evidence attached.
Does not log → covers the prompt. Not the click. There is no technical guarantee, and this week the industry admitted it.
Every one of these moves the boundary. None produces the proof.
Sovereignty you cannot audit is not sovereignty.
trustlayer.itmethods.com/p/t…
1
1
43
Your agents are shipping code. Who is proving it was allowed?
The Dark Factory: agents build it, a human approves every merge, Reign proves it. Regulator-ready by default.
15s below → itmethods.com/dark-factory/p…
1
40
iTmethods retweeted
.@itmethods has joined the Linux Foundation's FINOS and Agentic AI Foundation initiatives to help advance governance standards for agentic AI in regulated industries.
See what's driving AI governance 👇
#AI #OpenSource #Governance
pureai.com/blogs/the-pure-ai…
2
1
1
45
Alex Karp gave enterprises five questions for every AI vendor.
Who owns the data. Where is it cached. Are the prompts secure. Who controls the weights. Is your alpha transferring.
He aimed them at the frontier labs. The worse answers sit further down the SaaS stack, where nobody is asking.
Our CEO Paul Goldman in The Trust Layer, with a vendor assessment that grades the answers:
trustlayer.itmethods.com/p/t…
30
An AI software engineer just hit 492 million dollars in revenue, with Goldman, Citi, and Santander as customers. Gartner says agents will write nearly half of enterprise code this year.
The money is going to the coding agents. Almost none is going to the supervisor that governs the floor and proves what they did.
Isolation is not governance. The software factory is going dark. In regulated industries, the winners will keep every receipt.
The Trust Layer: trustlayer.itmethods.com/p/t…
26
Verifiable proof of trustworthy AI just became an open standard. The Linux Foundation launched the Appia Foundation with Google, Microsoft, OpenAI, Mastercard, and others.
The part the agentic era still forces: proof that holds continuously, as agents act and models change.
Our CEO Paul Goldman in The Trust Layer: trustlayer.itmethods.com/p/t…
22
1/ Canada is building sovereign AI, and RBC just mapped the stack: models, compute, data centres, research. It is excellent, and one layer is still to build.
2/ No Canadian champion sits on the control and assurance layer above the model, the layer that makes sovereignty operational and provable. Last week's frontier-model recall showed why it is the whole game.
3/ The layer still to build is software, not steel.
Our founder Paul Goldman on the one layer Canada should least want to outsource:
trustlayer.itmethods.com/p/c…
10
1/ On Friday, the most capable AI model on the public market disappeared. Not throttled. Pulled off every customer on earth by a U.S. export-control order, not a vendor decision.
2/ Not one customer did anything wrong, and the model vanished from all of them anyway. You cannot govern, encrypt, or self-host your way around a model that no longer exists for you to call.
3/ Single-model dependency is a single point of failure.
The control layer outranks the model.
Our founder Paul Goldman on what the recall means for regulated institutions: trustlayer.itmethods.com/p/t…
10
