PodCompanions
Discussions: 2026-W43 — page 59
Discussions added each week. These are publication dates here; the original episodes may be older.
- 1:08:10–1:13:10
Marblestone offers uncertain timelines and discusses proposed cost reductions and funding needs for large-scale brain mapping.
- 42:43–46:50
Marblestone contrasts cached action values with planning and connects dopamine responses to reward-prediction errors.
- 1:26:10–1:30:45
Marblestone discusses automated mathematics and explains why a verified result still depends on choosing an appropriate specification.
- 46:50–49:25
The speakers discuss model-based planning and cultural knowledge as a possible analogy for learned shortcuts.
- 1:13:10–1:17:30
Marblestone discusses molecularly annotated optical maps, their functional limits and the need for technology development and funding.
- 1:44:54–1:49:54
Marblestone describes the Gap Map’s proposed research capabilities and discusses which kinds of projects might suit different institutions.
- 39:50–42:43
Marblestone discusses reinforcement learning in insects and contrasts learned associations with specialised circuitry for biologically relevant signals.
- 10:00–15:00
Marblestone describes proposed mechanisms for assigning learned significance to abstract concepts and emphasises the uncertainty of these neuroscience models.
- 1:17:30–1:21:40
Marblestone discusses organisational challenges in neuroscience and possible uses of brain activity as additional training information for AI.
- 5:00–10:00
Marblestone considers how inherited circuitry could connect bodily signals and social concerns to learned representations.
- 22:20–26:40
Marblestone considers sampling, learned inference and what copying digital systems might change about learning.
- 26:40–30:15
Patel and Marblestone discuss how inherited information might specify architectures and learning objectives without encoding all acquired knowledge.
- 0:00–5:00
Marblestone proposes specialised learning objectives and architectures, then discusses prediction as a possible description of cortical computation.
- 35:00–39:50
Marblestone considers evolutionary changes in cortex, social learning and the relationships among neuron numbers, wiring and cognition.
- 1:39:57–1:44:54
Marblestone acknowledges uncertainty about consciousness and considers memory consolidation, replay and possible biological analogues of attention.
- 55:00–59:52
Marblestone uses cerebellar-cell timing to question simple neural models, then distinguishes intelligence from alignment with human values.
- 30:15–35:00
Marblestone describes molecular cell-type diversity and discusses what it could reveal about specialised inherited brain structures.
- 1:21:40–1:26:10
Marblestone considers learning from neural activity before turning to proof systems, verifiable mathematical tasks and conjecture generation.
- 50:32–55:00
Marblestone compares biological and digital constraints, including energy use, copying, memory and the complexity of cellular machinery.
- 15:00–19:14
Marblestone discusses possible learning mechanisms and cites visual-cortex models as an example of biological constraints improving artificial systems.
- 30:32–35:08
Flowers compares recollections about a friend’s behaviour and movements, acknowledging possible innocent explanations and details the reporting could not establish.
- 40:10–43:43
Flowers discusses a late, unverified lead and limits on follow-up, then includes family members’ rejection of involvement and the absence of charges.
- 11:15–13:55
The hosts discuss uncertainty about the timing, limitations of surveillance and the reported disappearance of Amaya’s camera.
- 46:19–49:55
The hosts qualify a secondhand account and consider alternative explanations for the missing camera, acknowledging that its contents and relevance remain unknown.