IndividualAdditional person profiled
Richard Sutton
Individual researcher
Temporal-difference learning, Dyna, reinforcement learning, continual experience.
AI Thought Leadership · Individual
Ilya Sutskever / Safe Superintelligence
Recurrent language modeling, deep-network optimization, sequence-to-sequence learning, and scaled autoregressive models.
Type
Individual (Individual researcher and current organization)
Directory section
Additional person profiled
What to monitor
Primary technical publications and any future public technical artifacts from Safe Superintelligence.
Directory last reviewed
August 16, 2026
Why this source is monitored
Ilya Sutskever's lineage spans recurrent text generation, AlexNet, optimization through initialization and momentum, sequence-to-sequence learning, GPT-2, and GPT-3.
Topics. ML foundations; language
Official links
Reviewed topic lineages
Caveat and non-endorsement
Safe Superintelligence is Sutskever's current organization, but the reviewed record contains no comparable public SSI technical artifact. Do not transfer claims from earlier publications to SSI.
Evidence boundary. Safe Superintelligence is Sutskever's current organization, but the reviewed record contains no comparable public SSI technical artifact. Do not transfer claims from earlier publications to SSI.
Inclusion in this directory is a monitoring decision by Implement Agentic. It is not an endorsement, a partnership, a client relationship, or a claim that this person or organization endorses us. Product claims and personal forecasts must be traced to technical evidence before they are used as fact.
How we weight this kind of source
- Primary research: papers, technical reports, system cards, benchmarks, official repositories, university research, or government evaluations.
- Open ecosystem: code, weights, datasets, model cards, evaluation tooling, and reproducibility work. Verify license and provenance separately.
- Expert synthesis: interpretation by a named researcher or practitioner. Link through to the underlying evidence.
- Market signal: startup, investor, or portfolio activity. Use to detect direction of interest, not as proof of capability.
- Publication/index: discovery infrastructure. A listing establishes publication or preprint status, not correctness.
Other sources in this section
IndividualAdditional person profiled
Geoffrey Hinton
Individual researcher
Backpropagation, distributed representations, deep learning; later AI-risk commentary.
IndividualAdditional person profiled
Yoshua Bengio
Individual researcher
Neural language modeling, representation learning, deep learning, scientific ML, safety.