An engineer who built language models for the U.S. intelligence community in 2014 is reportedly now targeting what the report calls "Big Token," though the available information does not specify what that term refers to or what form the effort will take.
Early report · AI model development
An engineer with intelligence and medical AI roots turns to "Big Token"
A builder of early language models for the U.S. intelligence community is reportedly targeting an effort called "Big Token." The term and the plan are not yet defined in the available material.
2014
Year the engineer built language models for the U.S. intelligence community
A few hundred
U.S. hospitals still running early cancer-detection AI the engineer built
Undefined
What "Big Token" means, its form, funding, and product details
The engineer's path, as summarized
Background milestones in order, ending at the open question
2014
Builds language models for U.S. intelligence
→
Earlier
Early cancer-detection AI now in a few hundred hospitals
→
Founded
DataFleets, described as training AI on confidential data
→
Now
Reportedly targeting "Big Token," details unconfirmed
What is not yet known
Treat this as an early, unverified outline rather than a confirmed development.
What "Big Token" refers to and what form the effort will take
Company name, product details, and funding figures for the new effort
The technical method behind DataFleets' training on confidential data (the summary is cut off)
Safeguards on data access and model retention tied to current plans
Independent confirmation of the engineer's history beyond the summary
Why it matters: If the effort is clarified, it could show how experienced builders from defense and healthcare AI are positioning themselves for data-sensitive AI systems.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article...