Can our data center expertise help with AI verification?

(Also posted on EA Forum.)

Treaties to pause training or otherwise restrict dangerous uses of AI would require difficult verification schemes. Plan A's proposal is that network taps inside data centers would record copies of traffic (or hashes of it), and later recomputations on random samples could check whether sampled outputs match declared workloads, as part of a broader verification scheme.

I'm an AI engineer at a company building tools for the construction industry. Much of our work is with data centers. The architects, engineers and construction managers on our team work every week with data center developers, their design engineers, and the general contractors building their sites.

As of now, the company does nothing at all involving AI verification, and I'm writing this in my personal capacity to try to learn: Can we be useful (e.g. helping with research) in these efforts? I do not speak for the company, but my hope is that if I learn how we can be useful, I could use my influence to make a case for this becoming a company-endorsed effort (albeit a side project).

Our existing work

We have a few features:

• Find design problems in construction documents
• Compare subsequent versions to previous, and see what issues may be resolved
• Review submittals (drawings or product data submitted by contractors to the design team) against each other and the design documents
• Review RFIs (requests for information from the contractor to the design team)
• Review proposed equipment purchases against a basis of design
• Tie all of the above into the construction schedule, e.g. identifying when design issues must be resolved to avoid impacting the planned activity or critical path
• … and more

This is all built via AI engineers like me working closely with domain experts like architects, electrical engineers, construction management experts, etc.

What the plans require building

Some of the writing on this that I've looked at over the last few days are:

• Plan A's verification plan
• Amodo's SITREP on it
• Cankaya's system overview and tap note

(But: There's a lot! A lot of work has gone into this and I just started thinking about it so despite reading those, I still haven't internalized much of it.)

The tap-based approaches require physical changes to what goes into a new data center as well as retrofitting existing ones:

• Network taps on fiber links, passive optical splitters where possible
• Physical defenses to prevent bypassing taps
• Recomputation inside the data center being monitored, trusted by the verifier but "possibly under [the prover's] physical control" (Amodo)

Plan A's retrofit scenario assumes most of the equipment is powered down during the work and includes a one-month retrofit target:

> retrofitting 40% of the US compute into inference units of 4 server racks (or 4K H100e) with their existing stock of network taps in 1 month

I didn't find a labor or installation-sequencing estimate behind that one-month timeline.

Where we could help

Amodo's SITREP explicitly calls for an installation playbook and says there has been little work on how to inspect verification systems at scale. Those sound like areas where construction expertise might help.

Any work we do on this would draw on our people's general professional expertise, on public reference designs, and on the developers and contractors we work with. I think it's possible that (with effort) our team (and maybe product) could help with:

• What's the particular sequence of construction activities involved in retrofitting a data center, including any work needed around equipment that remains in operation?
• Are there any design problems with the proposals, from a construction or operations standpoint?
• What would the "spec" for a new verification-ready data center look like?
• What do inspectors need to do to confirm all of this (taps, side channel defenses, networking, enclosures) are installed correctly? (This looks like a commissioning problem, which construction already has a process for. Maybe some of that can be borrowed.)
• How does the trend toward prefab components affect the feasibility and timing of these changes?

I don't know what form our help would take, e.g.:

• Just consultations
• Research artifacts produced by us (with advice from the folks working more on verification than we are!)

Should we talk?

Our company will not shift its focus to these problems. For now, I’m looking for people working on verification who see a specific use for this expertise. If there’s a fit, I could explore whether we can help. I’m not sure whether the teams pushing verification forward already have access to the kinds of experts I work with.

We don't directly build or operate data centers ourselves, but maybe this isn't too far from Plan A's Get Involved page requesting help from "datacenter builders and operators who know what is operationally feasible."

If you're working on these problems and think a small portion of our time could be useful to you, please comment or reach out to me. I'd also appreciate introductions to people working on installation and inspection.