Private academic seminar notes on Mac

Academic seminars, guest colloquia, and department symposiums represent the intellectual vanguard of scientific discovery and humanities scholarship. In these intensive intellectual gatherings, visiting professors present pre-publication manuscripts, doctoral candidates defend dissertation methodologies, and researchers debate competing interpretations of newly gathered empirical data.

Yet capturing the rapid intellectual exchange of academic colloquia is notoriously difficult:

Presenters move swiftly across dense slide decks, presenting advanced mathematical formulas, novel biochemical pathways, or intricate philosophical arguments. Expecting researchers or graduate students to type verbatim lecture notes while actively digesting complex theoretical models is impossible. However, introducing commercial cloud recording bots into departmental colloquia introduces unacceptable intellectual property liabilities: scholars routinely share unpublished laboratory findings, unreleased grant proposals, and proprietary patent disclosures that must never be ingested by third-party cloud AI vendors.

Oyma resolves this academic challenge on macOS. Leveraging Apple’s native ScreenCaptureKit framework, Oyma captures the audio of academic seminars silently without introducing an intrusive AI bot into your university Zoom, Microsoft Teams, or Webex meeting. Transcription executes 100% on-device on Apple Silicon using local Whisper models, transforming spoken academic debates into structured, permanent literature notes in plain Markdown stored securely in your local research vault.

The unique documentation demands of academic research

Scholarly note-taking requires synthesizing complex conceptual models while preserving academic integrity:

1. High cognitive listening bandwidth: Absorbing an advanced lecture on quantum error correction or macroeconomic econometric models demands undivided intellectual attention. Taking rapid physical notes detracts from conceptual understanding.
2. Pre-publication intellectual property exposure: Academic priority and patent filings depend on confidentiality prior to formal peer-reviewed publication. Exposing colloquium discussions of unpublished data to multi-tenant cloud transcription services violates university research ethics and grant requirements.
3. Complex academic jargon and non-standard nomenclature: Generic speech-to-text algorithms routinely mangle specialized terminology, turning technical acronyms and Latin species names into incomprehensible gibberish.
4. Disconnection from academic citation networks: Lecture notes taken in ephemeral notebook apps remain isolated from literature review bibliographies and ongoing dissertation drafts.

The scholar’s literature vault in plain Markdown

Oyma structures academic colloquium records directly into an interconnected Markdown knowledge base on your local disk:

vault/
├── research/
│   ├── seminars/
│   │   ├── 2026-10-02-mit-quantum-colloquium.md
│   │   ├── 2026-10-05-cern-lattice-symposium.md
│   │   └── 2026-10-09-stanford-neuro-lecture.md
│   └── literature/
│       ├── @vaswani2017attention.md
│       └── @lecun2022worldmodels.md
└── dissertation/
    ├── chapter-02-literature-review.md
    └── chapter-04-methodology.md

Connecting seminar debriefs to formal literature notes and thesis drafts via [[wikilinks]] enables researchers to build an expansive Zettelkasten knowledge network.

Sample on-device academic seminar note

Here is an authentic Markdown artifact generated during an academic colloquium in Oyma:

---
colloquiumTitle: "Non-Equilibrium Thermodynamics in Biological Neural Ensembles"
guestSpeaker: "Dr. Alistair Finch (Cambridge University)"
hostDepartment: "Department of Computational Neuroscience"
date: 2026-10-09
duration: 56m
citationKey: "@finch2026nonequilibrium"
relatedLiterature:
  - "[[research/literature/@friston2019freeenergy]]"
  - "[[research/literature/@hopfield1982neural]]"
tags:
  - academic-seminar
  - computational-neuroscience
  - thermodynamics
---

# Academic Seminar: Dr. Alistair Finch (Biological Neural Ensembles)

## 1. Abstract & Theoretical Premise
Dr. Finch presented pre-publication empirical findings exploring energy dissipation constraints in mammalian cortical microcircuits. The lecture demonstrated that dendritic computation operates near thermodynamic Landauer limits during burst-firing states, challenging classical feedforward energy models.

## 2. Core Hypotheses & Experimental Findings
- **Energy Dissipation Measurement:** Using sub-millisecond two-photon calcium imaging combined with micro-calorimetry, the Finch Lab measured local heat dissipation across 4,000 cortical pyramidal neurons in vitro.
- **Key Empirical Revelation:** Synaptic vesicle recycling accounts for 68% of metabolic expenditure during high-frequency gamma oscillations, whereas action potential propagation consumes only 22%.
- **Speaker Theoretical Formulation:**
  > "Biological neural networks do not optimize for raw computational throughput; they optimize for information transmission per unit of thermodynamic entropy generated."

## 3. Theoretical Debates & Q&A Synthesis
| Inquiring Scholar | Theoretical Challenge Raised | Speaker Response & Resolution |
| :--- | :--- | :--- |
| Prof. Marcus Vance | Incompatibility with standard Free Energy Principle models | Reconciled by demonstrating that active inference minimizes variational free energy subject to strict metabolic ceilings |
| Dr. Elena Rostova | Potential artifact introduced by in vitro perfusate temperature | Confirmed findings were replicated in vivo across awake rodent cohorts with zero temperature artifact |

## 4. Connections to Dissertation & Open Hypotheses
- **Thesis Impact:** Finch's empirical vesicle recycling measurements provide direct physical validation for the energy-bounded plasticity equation proposed in [[dissertation/chapter-04-methodology]].
- **Open Hypothesis:** Investigate whether artificial spiking neural networks can adopt Finch's Landauer dissipation regularization penalty to reduce GPU inference power consumption.

## 5. Follow-Up Reading & Research Tasks
| Priority | Action Item | Target Date | Status |
| :--- | :--- | :--- | :--- |
| P1 | Review Finch Lab pre-print on arXiv once released | 2026-10-16 | Open |
| P2 | Cross-reference mathematical derivations with [[research/literature/@friston2019freeenergy]] | 2026-10-20 | Open |
| P3 | Run simulation script testing vesicle recycling constraints on local PyTorch model | 2026-10-28 | Open |

The scholarly synthesis workflow: from seminar lecture to research paper

Transforming spoken colloquia into durable scholarly citations follows a rigorous, privacy-respecting protocol:

1. Silent seminar audio capture

When joining a virtual department colloquium on Zoom, Microsoft Teams, or Webex, engage Oyma‘s background audio listener. Utilizing macOS ScreenCaptureKit at the system tier, recording proceeds quietly. Without an automated attendee bot announcing itself, visiting scholars present unreleased pre-prints and preliminary findings without fear of surveillance.

2. Specialized academic acoustic recognition

Dense scholarly dialogue—including Greek mathematical notation, biochemical enzyme names, and theoretical physics terminology—undergoes immediate speech recognition via local Whisper models on Apple Silicon. Audio data remains strictly on your Mac’s encrypted hardware.

3. Synthesizing theoretical models and seminar debates

Following the lecture, execute Oyma‘s academic seminar extraction prompt using local Ollama models (such as Llama 3) or your private API key. The summary engine parses:

• Core scientific hypotheses and novel experimental findings
• Mathematical formulations and thermodynamic or econometric principles
• Vigorous debates and Q&A exchanges between faculty and the guest speaker
• Concrete connections to ongoing dissertation research and literature citations

4. Direct integration into your research Zettelkasten

The structured Markdown seminar note saves directly into your local research repository. Through bidirectional wikilinks, insights link to peer-reviewed literature notes, author profiles, and thesis chapters in Obsidian.

Protecting pre-publication intellectual property and academic priority

Academic research requires stringent confidentiality to protect scientific priority and patent rights:

• Preservation of academic priority: Scholarly discoveries rely on precedence. Storing seminar notes exclusively on local encrypted storage ensures that confidential laboratory findings are never leaked prior to formal publication.
• Air-gapped research: With Local-only mode, scholars can transcribe symposium audio, synthesize research notes, and explore citation graphs in field stations or air-gapped laboratory environments.
• Permanent, searchable research archive: Unlike cloud tools that lock your academic knowledge behind recurring fees, your local Markdown seminar notes remain permanent, accessible files that will survive long-term operating system updates.

To learn more about our local audio architecture, read about botless meeting recording, explore transcribing audio files, or review our zero-leak privacy architecture.