Running Notes

With: Jacob Ramirez (Forssell-Ramirez Makery seed grant) Tags: [CHECK-IN], [ADVISE]

Key Points

• Narrowing the scope of the bot — making it a specific part of the process rather than trying to cover the whole arc.

(Notes captured live; thin pass — flesh out if needed from memory or follow-up.)

Tags: [STUDIO-BUILD], [ALIGN] Source: Reuben’s captured thinking (likely post-conversations with Josh, Dan, Darcy)

Context

Pricing sheet (Flash Lab cost sheet) is in Josh’s hands. Broader question surfaced: AS needs external-facing collateral. Darcy called what we need a “tear sheet”; Dan asked what a one-pager for the studio would look like. Two distinct artifacts emerged.

Two artifacts

1. AS One-pager (who we are) — based on the Accelerator Studio powerpoint

• Side 1: Purpose, value, what we do concretely (specializations + services)
• Side 2: How you can plug in
• “Platform” framing (Dan’s word): make it clear how anyone can hook in

2. Services tear sheet (2-pager / front-and-back)

• Side 1: What we do — Flash Lab, Skill Sessions, Learning Impact Exchange
• Side 2: How you can plug in / get involved
• Quick-and-dirty outline first; pare content down hard

Audience (both documents — external)

• Funders
• Potential partners at organizational level
• District PD — looking for AI offerings, not necessarily school-level change but activating communities to think in new ways
• Cross-pollination: setting the enabling conditions for people to unlock and share their own expertise. “Activating.” Setting the stage for great conversations to follow.

Decisions / direction

• Co-design the tear sheet with Darcy
• First draft by sometime next week
• Have one ready well before ISTE — and before our meeting with Joseph South
• Today: send Darcy a WIP note: “Dan mentioned it would be nice to have an overview one-pager — we’re working on a 2-pager; what would be helpful to you as a development officer?” cc Josh

Action Items

• [ ] Email Darcy today (WIP note, ask what’s helpful, cc Josh)

  Owner: Reuben

  Due: 2026-05-08

• [ ] Draft AS one-pager (who we are)

  Owner: Reuben

  Due: next week, before ISTE/Joseph

• [ ] Draft AS services tear sheet (2-pager)

  Owner: Reuben

  Due: next week, before ISTE/Joseph

Source: Quick capture (likely Josh tip) Summary: Lakshmi (in Chris Lemons’ lab) wants to do a Flash Lab over the summer. Josh will make the intro. No further detail captured.

Action: Wait for Josh’s intro; add Lakshmi to people/ once name is confirmed.

With: Josh Weiss, Joe Sherman Tags: [CHECK-IN], [ALIGN], [STUDIO-BUILD]

Key Points

Josh’s Updates

• ThingLink legal concern: New T&C grants ThingLink extensive rights to student work — Stanford legal not happy. Rachel and Dan are working on it; resolution TBD
• New FTE: Josh has a job description in progress for an additional team member; passed HR review with multiple people so far. More to come
• Ed Ethics at GSC: Myra Levinson (new from Harvard) starting an “education ethics” group — framing is different from traditional ethics. Check out the Schools In podcast episode. Worth knowing about early — buzz is building
• School of Medicine EdTech partnership: Met with their group; co-proposal coming for a Q-Centered Therapy trauma intervention (ages 8–18, schools context). Kristen Bassetti is shopping for partners. Josh’s position: SAL won’t run the whole project (no capacity), but can contribute learning design expertise for school-age kids/school environments — SoM EdTech can handle project management. They have $25M+ and are figuring out scope. Co-proposal arriving today; Josh will share with SAL leadership
• Flash Lab demand increasing: Josh filtering incoming asks using partnership criteria; may route some to Reuben

Flash Lab Partnership Criteria (documented in discussion)

1. Genuinely excited about the core offering — not Stanford prestige or AI trendiness
2. Skin in the game (usually financial, but other forms count)
3. (Third criterion not recalled — find the source doc)
   • Train-the-trainer preferred; direct facilitation OK if it strategically builds a relationship or tests a prospect

Reuben’s Updates

• Tinkery Flash Lab proof-of-concept: Greg facilitated independently yesterday (May 4) with a new group after co-facilitating ~a month ago. Model is working ✅
• ISTE Live blurb finalized: ~20 revisions with Josh. 139-word version tested on non-technical audiences. Key framing: Designer / Builder / Advocate roles. Josh’s edits removed jargon (e.g., “problem statements,” “colleagues”); Reuben’s wife “this makes sense” = green light
• AI developer conference validation: Speaker at conference independently described hiring criteria as designer / builder / advocate (different words, same model) — strong external validation of Flash Lab’s framework
• Distinguished Careers Institute Flash Lab — June 5: Originally planned for 150 people, now split into ~50-person groups. Josh’s family vacation moved, so he’ll be there. Reuben taking June 8 off (anniversary)
• Build-a-Bot now multilingual: German (native speaker confirmed), Spanish, French, English live. Chinese pending native speaker verification. ~1,000 bots on the site. Delivered ~90-min virtual workshop for META group (Christine, Victor + Germany cohort). Activity still relevant ~2.5 years in — proof that shelf-life is baked into the design
• Flash Lab vs. Build-a-Bot distinction (working articulation): Build-a-Bot = “Is this problem bot-able?” — under the hood, foundational fluency, you already have a solution in mind. Flash Lab = “What’s the balance of AI vs. human in this R&D process?” — widening aperture on the problem space, not assuming a bot is the answer. Credit to Nancy for “bot-able.”
• Lane (Las Vegas school director): Potential Flash Lab customer via Christine’s referral. Intro call coming up

Joe’s Updates

• Early Careers event highlight reel live on YouTube; positive reception
• Our Voice project: nearly final — faculty edits incorporated, social media version with Vicky almost done. Isabel Sacks (SAL comms) doing a Joyful Learning grant feature + Q&A with Bridget; will incorporate Our Voice media. Reuben connected Bridget to the project; waiting on her feedback
• Ethnic studies grant videos (~5 individual project profiles): nearly finalized, approved by Maisha/Dan/Isabel. New request: add a short intro video (~1–2 min) explaining the seed grant itself, so videos have standalone context. Working with Mispa to schedule Maisha (+maybe Entero) recording. Finn will also cut a punchy social media compilation reel from all five interview sets
• Sea Grant story postponed — participant had a death in the family. Rescheduling with sensitivity
• VFT teaser videos: working with Kyla and Kristen, reusing existing experience footage for YouTube → website traffic
• HarmonAI Sea Grant story: Scheduling with Elizabeth for next week at Bing. Plan to get B-roll of students with Leslie + Reuben’s Cheddar model
• Learning Impact Exchange pilot (next Thursday, May 14): 15 RSVPs so far (target 10–25). Working with Sarah E on logistics. Reminder going out this Thursday. Josh adding 2 invitees (industry-to-ed-partnership person; teacher from yesterday’s Tinkery Flash Lab). Food ordering TBD (see action items). Form/link shared via Slack
• Checked in with Sea Grant projects ahead of Kathy meeting next week. Ryan (completed her project) has findings on parent involvement; wants to stay connected to future convenings. Most grants applied for extensions; next event likely fall

Tomorrow’s Meetings (May 6)

• Jalen & Roy / “Learning Through Creation”: First meeting, set up by Marama. Exploratory — they want help with learning design and research. Not currently in the tracking doc. Reuben to reach out to Kathy for context
• CCT sync: Joe sent Reuben the console notes on Slack. Josh forwarding Claire’s email/co-proposal to Reuben

Decisions

• SAL will not serve as project management lead for the SoM / Q-Centered Therapy project; offer learning design and school-environment expertise as scoped contribution
• Flash Lab facilitation priority: train-the-trainer first; direct facilitation only when strategically valuable
• Build-a-Bot multilingual rollout: confirm Chinese translation with native speaker before launch
• Learning Impact Exchange invitee approach: keep invite-only, use waitlist if needed, send Thursday reminder

Action Items

• [ ] Reach out to Kathy for context on Jalen & Roy “Learning Through Creation” meeting tomorrow

  Owner: Reuben

  Due: Today (May 5)

• [ ] Review CCT console notes Joe sent on Slack

  Owner: Reuben

  Due: Before tomorrow’s meeting

• [ ] Forward Claire’s email/co-proposal to Reuben

  Owner: Josh

  Due: Today

• [ ] Share Learning Impact Exchange form link in Slack

  Owner: Joe

  Due: Today

• [ ] Contact Leslie (or Alyssa) about P-card process for event food

  Owner: Joe

  Due: This week

• [ ] Send Learning Impact Exchange reminder to network + consider Joyful Learning contacts

  Owner: Joe

  Due: Thursday May 8

• [ ] Add Josh’s 2 invitees to Learning Impact Exchange

  Owner: Joe + Josh

  Due: This week

• [ ] Get co-proposal from SoM / Kristen Bassetti; share with SAL leadership

  Owner: Josh

  Due: This week

• [ ] Confirm Chinese translation with native speaker for Build-a-Bot

  Owner: Reuben

  Due: TBD

• [ ] Find the Flash Lab partnership criteria doc (recover the 3rd criterion)

  Owner: Reuben

  Due: TBD

• [ ] Schedule Maisha (+Entero?) recording for ethnic studies intro video

  Owner: Joe + Mispa

  Due: Next 2–4 weeks

• [ ] Reschedule postponed Sea Grant story (participant had death in family)

  Owner: Joe

  Due: TBD, give time

• [ ] Confirm HarmonAI / Elizabeth recording dates for next week at Bing

  Owner: Joe

  Due: ASAP

• [ ] Ensure Ryan (completed Sea Grant) is included in future convenings

  Owner: Joe

  Due: Before fall event

With: Sergio Andres Arango, Micaela Bonilla, Reuben Thiessen Format: Zoom · ~60 min · Continuation of Apr 17 collaboration. Joe Sherman OOO. Tags: [CO-CREATE], [TOOLBOX]

Key Points

• Prompt Lab language strategy — open question: should the team prompt in English and have the LLM respond in Spanish, then localize to Colombian Spanish? Discussed quality tradeoffs.
• Conversation quality degradation — Sergio/Micaela observing that the quality of conversation gets worse over time. Working memory in Lovable is part of the issue.
• Model switch — agreed to stay on Lovable but switch model to Gemini 3.1-flash. Team has budget to use whatever Reuben thinks is best.
• Data + risk — how risky is the current data setup? Reuben to draft a security document and send to the team.
• Funding the upgrade — Sergio to ask Wilson about getting a Gemini API key using funds from their seed grant.

Action Items

• Reuben: Draft security document for SofIA data setup; send to Sergio + Micaela
• Sergio: Ask Wilson re: Gemini API key from seed grant funds
• Sergio/Micaela: Test the Lovable + Gemini 3.1-flash switch and report back at next session (Thu May 7)

With: Josh Weiss, Joe Sherman, Reuben Thiessen Format: Zoom · ~50 min · End-of-month abbreviated version of the four-question rundown

Roles

• Joe: “What did we do?”
• Reuben: “What trends are popping up?”
• Josh: “Stories” + “What should we change?”

What did we do (highlights)

• Joe: Closed Shaping the Narrative workshop and the Early Careers reel (going up today). Ethnic studies videos near complete; Our Voice video done pending approval (social-cut version in flight w/ Vicky + Finn). Prepping the next round of seed grant story shoots starting next week.
• Reuben: Two Flash Lab implementations (Tinkery + outside group); the Lake Forest revisions are landing well. Trained Gregory on the post–Lake Forest changes; Greg’s solo Flash Lab moved from Apr 17 → May. Used Joe’s Shaping the Narrative workshop as a connection catalyst for the four seed-grantee CHECK-IN emails — sparked the Cyan meeting. Two new seed-grant onboardings: CRISPRkit (consulting Belinda on Android paths) and SofIA (deeper engagement; got source-code access).
• Joe (Sea Grant side): Designed a user-persona template + AI prompt for SofIA so the team can spin up agents from a filled-in persona image — useful both for student personas and for teacher-PD personas (practice working with a difficult vs. high-achieving student).
• Josh: Twice presented the data-viz “bubbles” deck (small SAL group + leadership). Vibe-coded the visualizations himself — landed well, generated “you don’t need a data analysis team” energy. Calling out partnerships moving slower than hoped (ISTE, Flash Lab beyond Stanford). Has shifted personal cadence to weekly Do-Not-Disturb strategy block (was monthly).

Trends

• Cross-team support catalysts. Joe’s workshop became Reuben’s vehicle for low-touch seed-grantee re-engagement; this is a pattern we should keep using deliberately.
• Breadth from a team of three. B-roll/stock photos → persona design → technical/architecture advice. Surprising how much surface area three people cover.
• Mature vs. newly-formed seed-grant teams need different support. Mystery Machine (Doorley/Britos) and Jacob’s team mostly need sounding-board energy; SofIA and other new teams need hands-on consultation. This maps onto our high/medium/low-touch language but isn’t formalized.
• Vibe-coded MVPs as an emerging pattern. SofIA “got to something that looks good but doesn’t really understand how/why it works.” Reuben sees this as a candidate for a formalized seed-grant offering: helping vibe-coded prototypes get to a maintainable next step.
• Faculty directors much more engaged than before. Myesha + Phil were notably engaged in the studio presentations; quoted impact from Myesha. Other faculty don’t quite stick around — they don’t undervalue the work, just don’t see it as something they need to think about.
• Headroom management is working. When asks come in, the team self-calibrates and does a light check-in with Josh rather than needing him to gate the language.
• Team growth. New seed grants + likely new headcount soon → start shaping a job description.
• “Must-have” framing landing externally. Elizabeth Schumann and others now naming Accelerator Studio as essential service rather than a nice-to-have.

Decisions

• Add a “high/medium/low-touch” calibration to seed-grant support. Set initially after first meeting; recategorize as the relationship evolves. Joe framed it as a “support-o-meter.”
• Strategic Huddles get a new first bullet under “Big Picture Reflection”: Calibrate Capacity. Three driving questions:
  1. Has the level of support for individual seed grants stayed consistent?
  2. Are there any that have gone from high touch → low touch (or vice versa)?
  3. Do we anticipate any upcoming changes at the individual seed-grant level that will affect overall capacity?
• Strategic Huddle scheduled May 12, 2026, 2:00–3:30 PM with Cathy. Josh sends invite.
• Confirmed Strategic Huddle cadence is quarterly (was previously written as “as-needed” in the handbook — to be corrected).

Stories worth capturing (Quoted Impact candidates — external only)

• Cyan DeVeaux (Apr 23): Variant of “so helpful to talk to somebody outside the project / have a sounding board.”
• Mystery Machine team (Doorley/Britos): Genuine enthusiasm for AS support despite being a team that “would be fine without us” — value of relationship building.
• Myesha + Phil: Engaged during studio presentations — moves them from “kind of understood what we did before” to “much stronger picture.”
• Elizabeth Schumann: Has named AS as “important / essential” from time to time.

Action items

• Josh: send Strategic Huddle invite for May 12 (Reuben, Joe, Cathy, Josh)
• Update team handbook: Strategic Huddle cadence “as-needed” → “quarterly”; add “Calibrate Capacity” as first bullet under Big Picture Reflection with the three driving questions
• Reuben: bring vibe-coded-MVP-support-as-a-service idea back as a Strategic Huddle agenda item
• Reuben: send ISTE Live session description (already on plate; flagged here to thread it to the studio narrative)

Tags

[ALIGN] [STUDIO-BUILD] [REFRAME]

With: Cyan DeVeaux (she/her), Josh Weiss, Joe Sherman, Reuben Thiessen Format: Zoom · 43 min Project: Dancing through Time — MR for Black dance history. Seed grant, Cyan is 4th-yr PhD at VHIL.

Source: Spellar transcription summary (webhook didn’t fire — pasted manually 2026-04-24).

Headline

Discussion focused on (a) recruiting an intergenerational participant pool beyond Stanford for the immersive VR dissertation study, and (b) development needs around dialogue management + captioning in Unity. The Studio’s role is connector + sounding board on study design.

Decisions

• Cyan reaches out to Kathy Kerns (Stanford lead librarian) first as recruitment / sounding-board entry point.
• Propose recurring group check-ins every 4–6 weeks for ongoing support.

Participant recruitment & external partners

• Goal: intergenerational sample (18+, including older participants) — not Stanford-students-only (would bias results).
• Partner venues floated: Stanford libraries (Kathy Kerns), community centers, museums — specifically Tech Interactive (San Jose) where Reuben/Joe have a contact open to collaboration.
• Joyful Learning grant staff (Tim / Tammin) suggested as connectors to community-driven programs.
• Group agreed: hold outreach until the experience is cohesive enough to show partners. Need targeted screening/consent language for VR comfort.

Study design, measures & timeline

• Mixed-methods: motion capture + behavioral data + surveys + interviews, mapped to embodied-learning constructs.
• Funding through December; Cyan aiming for spring dissertation defense — build & data collection must align.
• Encouraged iterative design (multiple versions) and a power analysis once measures are locked.
• Cathy Chase (transcript said “Kathy Chase” — almost certainly Cathy) flagged as senior sounding board for study iteration.

Development needs & accessibility (Unity)

• Cyan needs in-VR dialogue management + live captions for participants.
• Suggested: consult Brian Brown and HCI collaborators for libraries / dev support; check Unity accessibility lessons.
• Studio offered to connect Cyan with developers who implemented captioning in past projects (e.g., the science lab demo).

Action items (filed to action-items.md)

• Connect Cyan with Kathy Kerns — Reuben to introduce
• Find & share Tech Interactive contact card — Reuben/Joe
• Connect Cyan with Brian Brown + HCI devs re: Unity captioning — Reuben
• Connect Cyan with Tim/Tammin (Joyful Learning) — Reuben
• Connect Cyan with Cathy Chase for design sounding board — Reuben
• Schedule recurring Cyan check-ins (every 4–6 weeks) — Reuben

Tags

[CHECK-IN] [BRIDGE-BUILD] [CO-CREATE] [ADVISE]

Spellar source link

With: Josh Weiss, Joe Sherman

Scheduling

• End-of-month pushed to Mon Apr 27, 10:30–11:15 AM (45-min experiment). Reuben out Tue/Wed at DeepLearning.AI Dev Day (Andrew Ng, SF, using Stap funds); Joe out Thu/Fri (parents visiting from Boston, Calistoga weekend).
• AS:DE moved to Mon Apr 27, 11:15 AM–12:00 PM — before Joes’s parents arrive Wednesday.

Josh’s presentation to SAL faculty directors (Apr 20)

Josh showed the vibe-coded Accelerator Studio data viz at the faculty directors’ quarterly retreat (Dan, Isabelle, Victor, Phil Fisher et al., ~half-day). Well received — second showing this week, both positive.

Headline numbers surfaced:

• 146 projects · 18 high-touch with seed grantees · 31 external partners · 70 external collaborators
• 170 critical boosts · 108 capacity-building instances · 173 cross-pollination instances

Drill-downs that landed:

• Tag cloud (seed-grant-only slice) — check-in, toolbox, bridge-build, sounding-board, nugget lead the volume.
• Per-project views (e.g., King: 5 check-ins, bridge-builds, vulnerable share, reframe).
• Macro MyBook view — “the only macro project” — shows full tag coverage.
• Impact quotes pulled from Kathy’s survey.

Reception: Faculty directors “stoked.” Phil Fisher hadn’t realized the breadth of seed grant offerings. Questions centered on methodology — how Studio decides what to engage with. Some appetite to give Studio more room/support. Josh thinks directors underestimate team size (feels like 3.5 people, they think 2).

Open threads from the viz

1. “Check-in” tag undersells the work. Reuben flagged it: a check-in is actually the relational/morale-building touchpoint that unlocks everything else, but the word sounds transactional. No resolution — keep thinking about rename or reframing.
2. Where does Joe’s video production work live? Currently not tracked cleanly — not a macro project, not a seed grant, but substantial ongoing effort. Candidate buckets: operations, a dedicated tag, or line-item entries without dates. Joe to noodle.
3. Stanford Flash / build-a-bot tracking is incomplete — co-created work with South Florida isn’t in the macro tracker; needs backfill.
4. Impact tags. If the new second-round grants fund scaling projects (not new ones), that’s the cohort where breakthrough/ripple/scale-moment tags should start appearing. Something to watch rather than force.

SAL news from the directors’ retreat

• New within-Stanford grant coming — managed by Jessica Tsang (former Dan grad student, co-author on his book, contract basis, taught a class Reuben took with Kristen).
• Second-round seed grants being planned — likely bigger dollar amounts (~$100K-ish), probably fewer projects, likely more ambitious/time-intensive per project. Could be scaling existing seed grants, new projects, or a combo. Details TBD.
• Capacity implication: larger grants = more per-project Studio hours. Worth watching whether Studio bandwidth holds.

Studio retreat — “Revenge of the Studio”

Targeting Thursday June 11 (full day). Avoids: week of June 8 (Reuben’s wedding anniversary), last week of June (Reuben vacation), ISTE June 29–July 1.

Action items

• Joe — finish long-form + social-media videos, plug into Impact Edge viz (new “disseminate” coverage)
• Joe — decide how to track individual video production projects (ops vs. tag vs. line items)
• Josh — keep the “impact tags” question warm; revisit when second-round grants take shape
• Team — keep thinking about the “check-in” tag label / framing
• Team — hold June 11 for Studio retreat (Revenge of the Studio)

With: Christine Bywater

Key Points:

• On Cohort 2 with Germans. They have one more Tinker Time.
• Whole group is people who teach teachers
• Tinker Times have built knowledge as they went along
• Cohort 2 is way more advanced than Cohort 1
• There are loud participants who know a lot
• They have gone deeper and faster. Last one was Teachable Machine — “how do we start thinking about the D in the UDC triangle?”
• Jessica Yani was going to do it but isn’t

What they want:

• “This is how you teach teachers under the hood”
• “You can build your own agents”
• They can only use Copilot to build in schools

Format notes:

• 90 minutes. Start with a connector.
• Current event connection
• Make a takeaways slide — putting things in short succinct takeaways or directions is helpful for this group

With: Belinda Yeung, Reuben

Key Points:

• Belinda wants to scale the app beyond current pilot
• App is built in React Native (not Swift — good news for cross-platform)
• Tech stack: Supabase (auth), FISE (vector storage), AWS, OpenAI (chat model, 4.1-nano)
• RAG is built on expected questions from the Nature paper
• They have an Apple Developer Account already
• Need Google Play Developer account — Reuben shared Play Console link
• TestFlight available: code NGEHQB

Discussion: Cross-Platform / Chromebook Path:

• React Native can target Android, which Chromebooks can run
• Need to figure out Android apps on Chromebooks workflow
• GSE IT may have Android devices / Chromebooks to loan for testing
• Reuben shared BrowserStack (https://www.browserstack.com/) for testing across devices

Action Items:

• [ ] Download CRISPRkit TestFlight (code: NGEHQB)

  Owner: Reuben

• [ ] Check with GSE IT about loaner Android devices / Chromebooks

  Owner: Reuben

• [ ] Research Android apps on Chromebooks for CRISPRkit deployment

  Owner: Reuben

• Belinda: set up Google Play Developer account (Reuben sent link)

Tags: [CO-CREATE] [ADVISE] [TOOLBOX]

With: Christine Bywater, Josh Weiss, Reuben

Schedule change: Isabelle added a Tinkery tour and Victor discussion to the Lake Forest visit. Workshop time reduced to:

• Tuesday 3:30–4:30 PM (1 hour)
• Wednesday 9:00–10:00 AM (1 hour)
• Flash Lab normally needs way more than 60 minutes — need to pivot.

Attendees: ~20 people. The visiting leader is a VIP.

Approach discussed:

• Could do the Activation Stage concept — 20-30 min Flash Lab intro + CTA for full experience
• Pick one aspect, do a good job on that, then talk about it
• “How are you thinking about this in your district?”
• They’ll have just come from the Tinkery — bridge from that experience
• Facilitate their thinking around AI in teaching and learning

Christine’s framing:

• Bridge: how do we use these tools in service of teaching and learning?
• Articulate the problem you want to solve
• Two angles: professional practice (e.g., bell schedule) or student challenge
• “What role do humans and AI play in developing these solutions?” — spiral back to this throughout
• “How do I make decisions about the task and the tool, and how do I maintain myself in the loop?”
• “It’s ok for tech to be the boost, but you need to tell us why you chose this tool, and how did it go?”

Tags: [ALIGN] [LEVEL-UP] [BRIDGE-BUILD]

With: Gregory Wilson II, Reuben

Key Details:

• Up to 16 students attending. Jenny Nadaner is York Scholars Director.
• Student tracks: STEM Education, Arts, Community Engagement, Global Studies
• Key question raised: How to make Flash Lab collaborative when all students are working on their own individual projects?
• Session plan still TBD: format, timing, flow (1 hour? longer to include capstone context?)
• Materials/logistics: what does the Tinkery need from Reuben? Laptops? Printed task cards?
• Role clarity: Josh wanted Reuben there in person for this first one. Gregory’s role (co-facilitating vs. observing) to be confirmed.
• York Scholars program: https://www.york.org/programs/york-scholars

Next steps:

• Finalize session plan and collaborative format
• Confirm materials and logistics with Gregory
• Prepare task cards for Build-a-Bot and Flash Lab

Tags: [CO-CREATE] [MULTIPLIER]

With: Priya Talreja, Marily Oppezzo, Michele Patel, Reuben

Key Topics:

Data Storage & Architecture

• Debated cloud vs local storage. Agreed to steer IRB toward cloud syncing as default with local-only as opt-out.
• Existing cloud-sync tied to Google ID. Conversations are timestamped and logged, but bot configurations/iterations are not tracked.
• Proposed adding a column to conversation logs recording which bot block/personality version was active per message.
• Raised concern about Firebase for health-adjacent data — explore medical school alternatives.

User Editing & Experiment Design

• Discussed allowing users to edit their bot vs. keeping it fixed. Hypothesis: editing increases agency, trust, and adherence.
• “Coaching your coach” — engaging, recursive narrative.
• Proposed A/B experiment: editable vs non-editable bots, measuring edit counts and behavior change correlation.

Moderation & Provenance

• Multi-stage moderation pipeline: keyword filter → OpenAI moderation endpoint → custom prompt (7 harmful categories).
• Adds latency but necessary for risk mitigation.
• Health responses need provenance — link to peer-reviewed sources, not commercial pages.

IRB, Consent & Documentation

• Present cloud-sync as default to IRB, retain local-only opt-out.
• Prepare consent materials covering storage, deletion, moderation.
• Meet with Shelley (IRB contact) — out until Mar 27.
• Possible modification of existing study consent for pilot/booster rollout.

Logistics

• Follow-up meeting being scheduled (Tuesday options discussed). Caitlin suggested as attendee.
• Upload docs and demo links; provision email access for demo accounts.

Decisions:

• Cloud syncing = default for IRB submission (local-only as opt-out)
• Track bot iteration/version metadata tied to each conversation row
• Meet with Shelley re: IRB requirements after Mar 27

Tags: [CO-CREATE] [ADVISE] [BOOST]

It sounds like there are several action steps and follow-ups planned after this meeting. Here’s a summary of the key points and next steps:

1. MOU/Research Practice Agreement:
   • Sergio mentioned needing a template for an MOU (Memorandum of Understanding) or research practice agreement with Cevail.
   • Nancy Waymack has experience in creating such agreements, especially with international partnerships involving schools.
2. Technical Consultation:
   • Reuben will connect with the team to discuss technical architecture and infrastructure questions related to hardening and fortifying the current prototype for robust user testing and research purposes.
3. Interaction Design & Persona Mapping:
   • Joe and Kathy are willing to provide guidance on interaction design, persona mapping, and validating personas through user testing.
   • They can also help with measuring learning outcomes and knowledge acquisition from interactions with teachable agents like Sophia.
4. Partnerships in Colombia:
   • Sergio is looking for potential partners in Colombia, particularly school districts or networks that are open to growth mindset interventions.
   • Anna Saavedra (connected through the lab) might have useful connections but is currently on maternity leave.
   • Sergio will help draft a brief request specifying interest in schools receptive to growth mindset initiatives.
5. Timeline for Follow-Up Meetings:
   • The team should aim for follow-up meetings within 1-2 weeks, depending on the urgency of technical consultations and design discussions.
   • For partnerships, it’s important to reach out promptly with detailed requests to gauge initial interest from potential partners in Colombia.
6. Next Steps:
   • Sergio will draft a brief request for potential Colombian partners and share it with Joe and Ruben for review before sending it out.
   • Reuben will schedule a technical consultation meeting within the next week.
   • Joe and Kathy will set up meetings to discuss interaction design, persona mapping, and learning outcomes measurement.

It sounds like you had a productive and comprehensive one-on-one meeting covering various aspects of your work and personal development. Here are some key points from our conversation:

1. Quarterly Goals: We finalized your quarterly goals and discussed how they could roll up into a session for the studio group in June or July.

2. Career Planning: You expressed interest in discussing career planning, particularly around changes coming up with GPS (Goal-Setting Process) in April and May. We agreed that it would be beneficial to have this conversation sometime in May before our retreat and planning sessions in June and July.

3. GPS Goal Setting: We will revisit the previous goal-setting process and discuss any hesitations or areas for growth, focusing on personal, career, salary, title, and steam (steam seems like a typo, possibly meant “team” or another context-specific term) growth.

4. Email List Management: You are working on creating an email list of seed grant winners to make it easier to invite appropriate people to workshops. We discussed the possibility of having separate lists for grant winners versus broader prospects who might be curious about our offerings.

5. Personal Development: You shared that you’ve been doing well with your fitness goals, losing weight and feeling better after incorporating fasting into your routine.

6. Upcoming Events: There are several events planned over the next month and a half as part of the cross-pollination push for the end of the year.

7. Future Meetings: We agreed to revisit career planning in May and continue discussing personal growth areas during our one-on-one meetings.

If you need any further assistance or have specific questions about these points, feel free to reach out!

With: Isabelle Hau, Tomás Jiménez (Professor of Sociology, Co-Director), Jasmine Dehghan (Deputy Director of Programs) Note: Josh told Reuben this is NOT the LAUSD meeting — it’s about the Institute for Advancing Just Societies. Claire Fisher Moffett (SAL) was invited but unclear if she attended. Tags: [BRIDGE-BUILD] [SPARK]

About the Institute

• Broad definition of race and ethnicity — caste, tribe, nation
• Challenges and opportunities associated with the US are global
• Want to do work in partnership with orgs outside the university
• Support faculty doing research

Key Initiative: Movement of People and Cities

• Harnessing opportunities and addressing challenges of people moving
• “Welcoming America” — outside partner for this initiative
• First big convening in May: 30-40 people
• Format: 4 faculty researchers + 4 practitioners giving matched presentations
• Three parts: (1) identify replicable bright spots, (2) identify areas needing more research, (3) big visioning session for what the group does together going forward
• Connected to SAL through Brian Brown — specifically for the “visioning” component
• Working with the d.school on design and facilitation

Jasmine’s Role

• Deputy Director, 10 months in
• Working on project management and broader collaborations

Advice from Tomás on Organizing Convenings

(Tomás has extensive experience organizing large convenings — 350-400 people, one livestreamed to 22,000 viewers)

Format that works well:

• 3-4 voices, 5-7 minutes each (researcher, practitioner, community member — equal time)
• These short talks “point the audience to think about the topic”
• Then scaffold the discussion: 2-3 focused questions for the group to discuss over the next hour
• “Super engaging, super intellectual candies” — people love this format
• Applied many times, “each time seems like a wild success”

Artifacts are key:

• White papers or similar written outputs are extremely valuable
• Requires assigning a dedicated writer (student, scientist writer, or external)
• “We have been able to have so much [impact]… people love when it comes out, you can share it”
• Serves as “a nice gel for the group” — shared ownership of the product
• Gets widely shared and quoted — amplifies impact tenfold

Video production:

• Short 1-minute video clips of speakers for social media
• “People like videos. It was a real unlock for us.”
• Useful as teasers for future events, sponsor packages, etc.
• They hire an external organization for video + social media photography on a quarterly contract
• Low cost but high impact

Student writers:

• Assign a student (paid or unpaid) to write a reflection blog post
• Good learning opportunity for the student, useful content for the institute
• Easy to do, low cost, students often volunteer
• Note: if it goes through IRB (e.g., formal research), it gets more complicated — easier if student writes independently

Engaging philanthropy:

• Partner closely with development/fundraising colleagues
• Develop target lists: existing funders + prospects
• Funders generally like attending — eager to learn, share perspective
• “VIP dinner” after event is a proven fundraiser: 10-15 people, reflections on the event, broader conversation — “each time we have a big check out of it”
• Some funders need high-touch invitation process (e.g., Gates Foundation has internal protocols)

KPIs for convenings:

• Diversity of groups represented
• People actually connecting (measured via survey)
• Actual collaboration outcomes (harder to measure but they try)

Handling sensitive topics (e.g., mental health, disability, refugee topics):

• Start with “meeting norms” (not Chatham House rules — since they do plan to publish something)
• Discuss norms at the start to ensure everyone is comfortable
• Don’t record sensitive sessions — no video
• Still plan for written artifact (white paper)
• Go back to every individual AND their organization for approval before quoting
• “A lot of work… but important to maintain trust and buy-in”

Cadence:

• 2 large convenings per year + smaller topic-specific ones (~50 people)
• Currently doing 3 this year because “AI is running us… it’s a moment to seize”
• Topics so far: Impact of AI on assessments, AI and ability/careers, AI and youth mental well-being

People Mentioned

• Maisha — recently recruited to Stanford, Tomás wants to connect with her
• Saadi (project manager at the Institute) — longtime friend of Tomás
• Tomás has a book coming out on refugee resettlement under Biden + community-based sponsorship
• He has a “secret weapon” — a former Stanford HAI event planner who now does freelance event management, still works with HAI. Inexpensive, knows Stanford well.

Next Steps

• Unclear what the capacity-building workshop ask is — need follow-up from Josh/Isabelle
• May connect to LAUSD Thursday meeting or may be a separate thread
• Tomás offered to stay connected for follow-up questions
• Potential areas of mutual interest to explore in future meeting

With: Rachel Wolf (AAALab) Tags: [SPARK] [THOUGHT-PARTNER]

Rachel’s New Project — AI in K-12 Ag Education

• Rachel has been part of a project for 10 years in Maine around climate science education
• Connected with 4-H people — separate thing funded by USDA around aquaponics education
• Kids in Maine — because of climate change, fishing is changing
• Applying for round 2 of existing grant + another grant “Strengthening Agriculture System” — using AI in K-12 ag education
• Gist: they want people using AI for K-12 ag education + workforce development
• Feels like a really cool space for learning by creating

Dream Scenario

• AI co-creation task: kids create something with AI, then compare what they generated with what’s in the real world
• Identify: innovations, imitation, what’s practical now, what’s practical eventually
• Problem-oriented, NOT making a chatbot — they see lots of chatbots and hands-on AI literacy programming
• What could it look like for kids to come up with ideas and use AI to compare and expand?
• See aquaponics in more of an abstract concept map

What Rachel Asked Reuben

• Recommendations for what to look at
• What’s possible now vs. on the horizon
• “Learning by creating” that is problem-oriented but not chatbot-building

Key Details

• 4-H has connections to National Aquaponics Association
• Kids’ favorite activity: tour aquaponics, then build with PVC pipes — hands-on learning
• Target age: upper elementary and middle school (under 13)

With: Josh Weiss, Joe Sherman Tags: [ALIGN] [STUDIO-BUILD]

Key Points

• Flash Lab: basic vs deluxe? — Discussion about what’s in/out of the toolkit. Need to create a document to inventory components (basic version vs deluxe offering).
• Health Coach Bot framing: If Reuben is going to be having ongoing conversations with Marily, what is the thing he wants to see happen that would evaluate the value of Build-a-Bot? Is Marily’s study going to help validate? This is a chance to test out what Reuben thinks the value of BaB is.
• Joe updates:
  • Elizabeth seed grant video: Joe connected Elizabeth with Vicki and Finn — working together on seed grant video. They are really pumped about it.
  • Marlow (Stanford UIT) reached out to HarmonAI team about using their computing resources to build the model.
  • Working on VFT stuff with Kayla + frame.io
  • Dan’s final advisory board meeting next week — group is working on a thank you video for Dan
• Feedback on Joe’s workshop — 10 min quick review (Google Doc: https://docs.google.com/document/d/1adVifJdOa2ximCoilIaBuvKUcbmBxHJuehOXZdHlSV8/edit?usp=sharing)
• Confidential: Stage 2 seed grant funding secured. Stage 1 average has been ~$50k; Dan has been working to get 6-figure funding. Small handful of projects will get Stage 2 funding and will need AS:DE support. Direction: “getting off the launch pad.”
• Crystal conference: Reuben to forward to Christina Hewko and Sarahi Espinoza Salamanca (DONE)
• Airtable: Note to get recent activities into Airtable

Action Items

• [ ] Create Flash Lab inventory document (basic vs deluxe components)

  Owner: Reuben

• [ ] Frame what BaB validation looks like for Health Coach Bot study

  Owner: Reuben

• [ ] Update Airtable with recent activities

  Owner: Reuben

• [x] Forward Crystal AI + Equity Conference to Christina + Sarahi

  Done: Mar 10

With: Isabelle Hau, Josh Weiss, Brooke (OTL — branding/trademarks), David (OTL — senior licensing) Tags: [ALIGN] [SCALE-MOMENT]

Context

Meeting with OTL to discuss IP and licensing structure for the ISTE partnership (train-the-trainer model for Flash Lab).

Key Points

• Business model: Train master trainers at ISTE/ASCD → they deliver to teachers → revenue share if teachers take the training
• YouCubed precedent: Developed summer camp curriculum, trained counselors worldwide. Branding: “Summer Camp inspired by YouCubed at Stanford.” Revenue comes from training the teachers. Brooke raised this as the closest parallel.
• Other precedents: Ed TPA (trainer-of-trainers model), Chariot program (Stanford Medical — did a nonprofit startup with royalty-free license)
• IP options discussed:
  • Open-source the day-to-day curriculum (creators can give it away per Stanford policy) → what people pay for is the training (simplifies to a services agreement)
  • Keep train-the-trainer materials proprietary → license those separately
  • Two-tier model: open-source workshop materials + proprietary trainer certification content
• Key insight from David: “Think about where the true value is — is it in the curated content itself, or in the training and skills you develop in the course?”
• Copyright: Anything created through your job with more than incidental use of Stanford resources = Stanford IP. But creators have the right to open-source.
• Participant creations: If someone creates something at a workshop, Stanford has no claim to it.
• Contract structure: Likely a Work-Service Agreement (WSA), not a formal IP license. Brooke thinks it’ll come down to some sort of WSA. Challenge: once they start doing training independently, how do we get cost recovery?
• Risk: ISTE could start developing their own similar training that’s not exact but competes.
• Revenue considerations: UBIC (Unrelated Business Income) has tax implications. OTL preference is to completely remove it from the university if possible — “that is the cleanest.”
• If no IP license: David noted “what teeth do we have?” — proprietary trainer materials would give leverage.
• Geoff Cox: Brooke consulted him — he’s developing a PL that would be facilitated through another org. May be the one to help structure this.

Decisions

• Two-tier approach is the likely path: open-source workshop materials + proprietary train-the-trainer content
• Services agreement (WSA) is more likely than formal IP license

Action Items

• Reuben: Add copyright notice and Terms of Use to designkit.stanford.edu
• Brooke: Follow up with Geoff Cox and with YouCubed for precedent details
• Team: Develop proprietary train-the-trainer materials (separate from open-source workshop content)

With: Josh Weiss, Joe Sherman Tags: [ALIGN] [STUDIO-BUILD]

Key Points

• Sarai wants to do a compressed “capacity building” for Create+AI people
• tEquity inter-cohort mingle: Josh feels there’s a more communal vibe among grantees. Sarai is OK taking the lead on an inter-cohort mingle but needs our backstage support
• Need to get sal-seedgrantee list access so we can maintain the grantee list for invitations
• Josh has been doing critical boosts — need to formalize that into our documenting process
• Reuben needs to update documentation for MyBook and HarmonAI with recent work
• Confidential: Dan is making moves for cross-Stanford education + AI. Thinking about a structured support pipeline like “build and experiment.”
• Josh traveling March 16-18 to U Mich Center of Academic Innovation
• Josh wants to discuss in next 1:1: “Where would you like to go in the next 12 months?” on the 3 C’s

Action Items

• Get sal-seedgrantee list access
• Update MyBook + HarmonAI documentation
• Follow up with Keith re: judging for Design Challenge (may conflict with Crystal)
• Prep for 1:1: 3 C’s / next 12 months conversation

With: Gregory Wilson (Tinkery)

Build-a-Bot facilitator training walkthrough. Gregory came away with a good understanding of the platform. Meeting went well.

With: Marily Oppezzo, Michele Patel Time: 9:00 AM Tags: [CO-CREATE] [ADVISE]

Key Points

• Reference paper with 93 behavior change techniques — use for bot content
• Need to add motivational content to the bot
• “About Me” section needs to be clearer
• Add RAG links to relevant documents when certain blocks are chosen
• “Exercise snacks” during the workday — “Snacktivity” = anything that gets your heart rate up
• A lot of people are struggling with ways to incorporate healthy activities into the workday
• Behavioral health coaching, not about prescription
• Data export discussion: can users just export their data?
• Confirm where data is stored — is it actually in Firebase?
• Consider anonymized summary reports

Next Steps

• Marily & Michele will send articles
• Reuben will build a Prompt Lab

Action Items

• [ ] Build a Prompt Lab for Health Coach Bot

  Owner: Reuben

• [ ] Confirm data storage location (Firebase?) for Health Coach Bot

  Owner: Reuben

• [ ] Add motivational content + incorporate 93 behavior change techniques

  Owner: Reuben

• [ ] Make “About Me” section clearer

  Owner: Reuben

• [ ] Add RAG links to relevant documents for chosen blocks

  Owner: Reuben

• [ ] Explore data export + anonymized summary reports

  Owner: Reuben

• [ ] Waiting: Marily & Michele to send articles

  Owner: Marily/Michele

With: Gregory Wilson II, Josh Weiss Time: 1:00 PM Tags: [CO-CREATE] [ALIGN] [BRIDGE-BUILD]

Key Points

• Flash Lab would be a great fit for a Tinkery space — Gregory enthusiastic
• Typically Tinkery workshops are one hour
• Similar goal with the AI Tinkery Toolkit — both toolkits serve educators getting started
• York Scholars opportunity: Group of junior and senior high school students from New York. They do a capstone — the goal would be to use Flash Lab to come up with prototypes for their capstone project. See: https://www.york.org/programs/york-scholars

Action Items

• [ ] Reuben to set a time to go through Build-a-Bot facilitator training with Gregory

  Owner: Reuben

• [ ] Making task cards for Build-a-Bot and Flash Lab (for use in Tinkery)

  Owner: Gregory/Reuben

• [ ] Set meeting for next week to do the overview facilitator training (Build-a-Bot + Tinkery Toolkit alignment)

  Owner: Reuben

• [ ] Pick March 30 for on-campus Flash Lab workshop at the Tinkery — confirm with Gregory

  Owner: Reuben

With: Christine Bywater Time: 11:26 AM – 12:00 PM Tags: [ALIGN] [CO-CREATE] [BRIDGE-BUILD]

Context

• Christine is very busy with PD. Goes to Germany March 8.
• Exploring how Accelerator Studio and CSET/PLEX can support each other
• Discussion of Flash Lab’s relationship to Build-a-Bot

Key Points

• CSET has stopped doing 1-day workshops — smallest learning experience is now 8 hours total across 2-3 days
• In a small series: teachers understand AI, understand AI literacy
• UDC triangle — most people think “critic” is critiquing the output, but it’s actually critiquing AI in society
• How do we slow down and look for bias?
• Critical consumer mindset → constructing habits that make sense for you → make decisions of when and why to use a tool
• Key phrase: “problem of practice” — Reuben to start using this language
• “In the loop” activity — pause at each point for reflection; examining routines and decision-making

Matthew Nelson (aside)

• Community college person Reuben has done some consulting with (trades)
• Couple of 1-hour conversations; can’t remember exact role

Lake Forest Opportunity

• Christine’s offering: Superintendent Matthew Montgomery, Lake Forest, Chicago
• Can bring teachers and admins to Stanford
• Monday March 30 – Wednesday April 1 — Christine will send schedule
• D-school workshop, Dan will talk to them — design thinking and AI literacy
• They have a chunk of time for Reuben’s content
• Proposed sequence: intro → Flash Lab → punch line
• Available time for Reuben: Tuesday March 31 starting 3:00 PM; free before noon Wednesday April 1

Christine’s 5-Session Series

• Christine has a 5-session sequence planned
• Flash Lab could fit in Session 3
• Working with org called “Ground Floor”
• Build-a-Bot might be part of it (online self-paced component)

ISTE Note

• Districts love their standards, have robust competencies
• When it comes to AI: they’ve started but make their own
• Complaint: existing AI standards/resources are too tool-centric
• AI becoming much more about environment, politics — districts don’t find that addressed in current offerings

With: Wen Profiri Time: 2:15 PM Tags: [CO-CREATE] [TOOLBOX]

MyBook Updates

• Dr. Wang will handle administration — already got kids from schools to sign up
• Three Palo Alto high schools, one already agreed to participate
• Changed framing: “caregiver” → “teen” — this is a 12-week program
• Data collection: weekly, through a 12-module system
• Survey: pre and post survey
• No medical information collected — basic surveys about AI trust and skills development
• IRB may not be necessary for simplified version, but Reuben suggested checking with Dr. Stafford to be certain
• Teens report data weekly; elders use Google sign-on, report through their teenagers
• Wen to send Reuben links and requirements for the dashboard
• Reuben will create a prototype and schedule follow-up for Friday

Next Steps

• Wen: Send links and requirements for dashboard
• Reuben: Create prototype dashboard for AI Trust Survey data collection
• Schedule follow-up meeting for Friday to review progress
• Check with Dr. Stafford on IRB necessity

With: Josh Weiss, Isabelle Hau, Reuben Time: ~3:45 PM (quick call right after ISTE meeting) Tags: [ALIGN] [THOUGHT-PARTNER]

Strategic Direction

• IP approach: Retain IP, license to ISTE. “Powered by Stanford” branding.
• Framing: Friends, not competition — additive to ISTE’s offerings
• Tinkery connection is relevant — feeds into the partnership ecosystem
• Outcomes: #1 focus = impact

Due Diligence Needed

• How many trainings does ISTE do? What is the revenue potential?
• Do due diligence on ISTE’s PL (professional learning) offerings
• Start as a pilot — small scale, see if the fit is still there

Open Questions from Debrief

• Would we be training the 90 contractors? Or offering Flash Lab as add-on modules?
• Are CLS faculty the same as the 90 contractors? (Need to clarify with Jessica)
• How would these trainings be delivered?
• How often would we have to conduct TtT? How transient are the contractors? Refreshers?
• They’re all meeting at ISTE Live — sign partnership before ISTE Live?

Quality & Branding

• More comfortable with us training ISTE staff first, then CLS faculty
• Quality control is a concern — what happens when we’re not in the room?
• Need to make sure branding is correct — “certified trainer” designation
• OTL: Isabelle has a contact — she’ll handle the OTL + dean conversation

Josh’s Written Reflections (email to Isabelle, 2:12 PM)

What feels right:

• Increases reach with practitioners without diluting the research-centered experiential learning that makes us unique
• Authorized provider model with IP retention sets us up for deploying with other authorized providers in the future
• Joseph has been great to work with

To think about:

• Tinkery involvement — Gregory has mentioned wanting quarterly CoP meetings with Tinkery partner schools, but getting that up and running might be challenging (hasn’t done it before). This could be a growth opportunity.
• Direct TtT only (for now): Josh is a fan of training ISTE staff directly. Less a fan of second-hand training (their staff training a trainer). Reuben or someone from SAL should have a direct hand in any TtT session, at least for now. Could explore a phase where they do TtT without us if things go well.

To keep an eye on:

• How educators perceive our authorized providers
• Conflict with other Stanford units doing AI PD (Reuben has been planning ways to address this)

Shared doc: https://docs.google.com/document/d/1mXa2fxnbLxEDsRfU8TfVapwgYENsbPTqXoqAvhbNgv4/edit?usp=sharing

With: Joseph South (ISTE), Jessica Garner (ISTE), Isabelle Hau, Josh Weiss, Reuben Time: 1:00 PM Tags: [ALIGN] [BRIDGE-BUILD] [SCALE-MOMENT]

Context

• ISTE just announced partnership with Google for AI literacy training to 6M US educators — they move fast (9 business days!)
• Link: https://iste.org/news/iste-ascd-and-google-partner-to-provide-ai-literacy-training-to-six-million-u-s-educators

Jessica — ISTE Offerings Update

• Lots of PL around AI for districts — Flash Lab could fit into this
• Kicking off communities of practice
• Second cohort (launched late Sept) was more challenging — how to build foundational knowledge for newcomers without holding back people ready to run
• Using AI to hone a problem of practice
• Kicking off next cohort — natural synergy with Flash Lab
• Asked: how could ISTE faculty get trained to use Flash Lab? What would that look like logistically?

Joseph — Conference & Training Logistics

• CLS (Custom Learning Services) faculty — most will be at their conference. Could do face-to-face training there.
• Activation Stage at conference: can’t run something longer than 1 hour. Also a stage for 20-30 min presentations.
• If there was a 1-hour version — could do something engaging that introduces Flash Lab
• Josh: could give a taste of what it is and what it produces — how it gets a teacher to think differently. Call to action: “let’s bring this full experience to my school/district”

Isabelle — Framing

• Two buckets evolving at SAL:
  1. Experiential learning — AI Tinkery model getting traction, lots in this bucket
  2. Academic research — where is AI working and for whom? Causal research still emerging
• Homework to do with OTL (Office of Technology Licensing), and need dean buy-in on collaboration/partnership

Joseph — Partnership Structure

• Experiential training: ISTE has the appetite to train their trainers. Just need to formalize the relationship.
• Academic research: Could run a series of articles in EL magazine translating knowledge into practice. If not the magazine, blog articles. Chris Agnew is a good follow-up contact for this.
• ISTE expanding their RAG tool (stretchAI) — could incorporate research insights
• Joseph (re: AI + research): “If only there were some people at your organization who knew something about how to use AI to contextualize research…”

Pricing / Business Model Discussion

• ISTE wants: training to do it right, use Stanford’s name. “Bring you this thing developed at Stanford, trained by Stanford to do it right.”
• Stanford side: “We’re looking to test out a compensation model for time and materials.”
• Joseph proposed: pay a fee to get trained, then share a fee, funnel participants into a Tinkery network
• CLS pricing: $7,500/day — all topics. Faculty are 95% contractors.
• Potential model: ISTE has teaching & learning offering at $7,500/day. Add Flash Lab for $X. They train, pay back rev-share.
• If funneling participants back to Stanford programs, rev-share would be lower
• IP question: Where does IP reside? ISTE not looking to take IP on things Stanford already developed — they want to be an authorized provider. IP stays at Stanford. Co-development would be a different conversation.
• Google funding angle: Joseph suggested Google could fund the differential — “Through the generosity of Google, we’re going to offer that Flash Lab experience.” ISTE as distribution channel.

Key Takeaway

Joseph essentially designed the business model for us. The authorized provider structure, the rev-share on $7,500/day CLS delivery, the Google funding angle — all of that came from him unprompted. He’s not just open to the partnership, he’s actively architecting how it works. This is a very strong signal of buy-in from the ISTE side.

Next Steps

• Isabelle: OTL conversation + dean buy-in on partnership
• Formalize the relationship — authorized provider model
• Explore Google funding angle
• Follow up with Chris Agnew re: research articles/stretchAI
• Figure out conference presence (Activation Stage 20-30 min intro + CTA)

With: Josh Weiss, Joe Sherman, Reuben Time: 11:00 AM Tags: [ALIGN] [THOUGHT-PARTNER]

ISTE Meeting Prep (for Wed Feb 25)

Framing: This is complementary for both sides. Both leaders (Isabelle + Joseph South) are in the room — let them talk, but know what we want.

If Joseph says “what do you want?”:

• Regular Flash Lab sessions
• By trained facilitators
• In as many places as possible
• BONUS: impact data on what came out of those sessions

Test within their bounds, then understand how it scales within their architecture.

Link back to mission: Reaching people aligned with seed grantees and GSE. Imagine Flash Lab outputs aligned with seed grantees (practice → research) — identify natural teacher-partners.

Three Personas Framework (NEW)

After someone goes through Flash Lab or Build-a-Bot, they land in one of three personas based on their driving question:

Design implication: “Today we’re going to explore three essential questions. At the end you’ll have one you want to continue on. If you’d like, we’re happy to connect you to these partners.” — Strategic, not random.

Key insight from Josh: “Do you want them to discover their value or affirm their value?” → Discover. Research doesn’t validate assumptions.

“Achieve Failure” concept: Confidently articulating what doesn’t work (so far). In the age of breathless AI headlines, it’s nice to have permission to find something it doesn’t work for.

Pricing Discussion

• “Testing a pricing model to understand the value”
• One hour + materials = $5k baseline
• Constant iterative process (ongoing partnership) = more
• If Joseph mentions short-term and long-term grant — this is the chance to talk $$ and need

Multiple Planes of Engagement

How does a teacher/parent/administrator/student’s journey look end-to-end?

Central questions each person is asking:

• “I’m hearing a lot about AI tools, but how could they specifically help me?”
• “How do I use these without shortcutting the learning process?”
• “How do I build something that takes advantage of AI but aligns with learning sciences?”
• “What’s my framework for whether or not to use AI?”

More implicit in workshop design: “Here’s how you can use this as an augmenter rather than an oracle.”

Action Items

• [ ] Think about pricing model in more detail (baseline: 1hr + materials = $5k)

  Owner: Reuben

  Due: Before ISTE meeting (Wed Feb 25)

  From: Feb 24 AS:DE

• [ ] Think about three core personas and which pathways map to which partners (CSET, Tinkery, etc.)

  Owner: Reuben

  Due: Next Josh conversation

  From: Feb 24 AS:DE

• [ ] For ISTE tomorrow: which of the three persona questions is most pertinent to their case?

  Owner: Reuben

  Due: Before ISTE meeting (Wed Feb 25)

  From: Feb 24 AS:DE

Next Steps

• Apply three personas lens to ISTE specifically for tomorrow
• Also apply to: Tinkery, CSET, international organizations
• Think about persona → partner pathway mapping broadly

With: Victor Lee, Christine Bywater, Victoria Delaney, Bernardo Silveira, Preetha Menon, Joba Adisa Time: 2:00 PM

Shane Update (via Christine)

• Shane (community college faculty) built a board with switches
• Given 15 random scenarios — tested for 3 days, worked well all 3 days

Protein Synthesis Lab Demo

• Built a protein synthesis lab simulation
• One person is the “thought partner” AI
• Flow: customizes conversation to the student → goes into the lab → debriefs
• In small groups: try to make 15 expert conversations
• Tech stack: All done in Unity; front end runs like a website with an OpenAI conversation layer. Program personalities.
• Prompt behavior: Everything gets a positive response. Goal: affirm, identify the right/wrong idea, then redirect.
• Challenge: AI was using way too many metaphors
• Upcoming RCT: May 5, Sonoma, 66 kids, two groups (one gets pen and paper control)
• Reuben’s contribution: Offered a quick fix on device orientation (Xcode) and offered Micah to test it

Facilitator: Sam McClure Format: Zoom, noon PT Organized by: Reuben (Accelerator Studio) for SAL seed grantees

Key Takeaways

SU-18 Agreement & IP:

• Everyone with a Stanford ID has signed SU-18 — Stanford owns IP created using Stanford resources
• If you do something off campus, not using resources, and it’s NOT in the scope of your research — it’s yours
• If it IS in the scope of your research — Stanford has a claim
• OTL (Office of Technology Licensing) is the resource for navigating this
• OTL has a form you can submit to them to start the process of clarifying IP ownership
• “Well-advised to have a conversation with an outside lawyer”
• Once OTL process happens, you can incorporate, etc. — this is the essential first step

Entity Structure & Incorporation:

• Sam has a playbook / white paper: https://ecopreneurship.stanford.edu/white-papers/academic-innovation-pathways/
• Questions about entity structure — when to start thinking about it? “When you’re getting serious”
• Having a mission is valuable for hiring, fundraising — powerful tool to bake mission into corporate documents
• Revenue model is to scale, not the end goal
• Fiscal sponsorship: a nonprofit lends its 501(c)(3) to you, but takes a cut of grants
• Hybrid (for-profit + nonprofit): more than twice as complicated to run both

Founder Dynamics:

• At least half of company failures in first 5 years fail because of ownership and governance decisions
• Sam discourages founder agreements — almost always leads to trouble. Reasonable for team/idea to change over time
• Professors with grad students: taking it out of lab into the world is a common path

Compliance:

• Should engage a lawyer for COPPA/FERPA compliance
• “Aren’t lawyers expensive?” — yes, but necessary when getting serious

For Reuben Personally

• SU-18 implications for Build-a-Bot, Flash Lab, and other tools built with Stanford resources
• Need to understand what’s in scope vs. out of scope
• Action: Submit OTL form to clarify IP ownership for key tools
• Action: Review personal files/projects on work laptop for SU-18 exposure

Location: ANKO, Shriram Hall II, Stanford Organizers: Stanford Accelerator for Learning + HAI

Key Themes & Quotes

Michael Taubman — “AI Driver’s License” framework: Four parts: (1) Choose the destination, (2) How to use?, (3) Open the hood, (4) Discuss/debate the rules of the road

On AI as tool AND topic:

• “Students need to learn about AI both as a tool and a topic”
• AI hasn’t sped up the thinking and collaboration (in a good way)
• “AI opens the door, the student walks through it”

On student awareness:

• Term heard: “AI autodidact” (self-taught)
• “Usually you have to convince students of the relevance of what you’re teaching. They are currently aware of the relevance — we are living through a historical moment together.”

On industry vs. classroom:

• “Industry is thinking about engagement differently and runs counter to what they’re looking at in the classroom”

Candace Thille — Responsible Assessment in the Age of AI:

• Standardization doesn’t truly assess what a student can do
• Constructs of fairness, comparability, validity and reliability
• Socially and culturally responsive in the age of AI
• Full white paper coming out in the next couple weeks

On technology and mission:

• “Anytime you introduce technology where technology is not on the mission, you amplify a bunch of not good stuff” (Teach for America founder)
• Used AI to develop seating charts — putting connector kids next to quiet kids

On coding costs:

• “Reduced the cost of coding to practically zero. So people who are trying to solve problems that can be solved by coding, can do things.”

Dan Schwartz on personalization:

• People think of personalization as: you have a curriculum and you move forward or back depending on where you are
• But what he’s seeing: people targeting demographics, making tools for markets the bigger market won’t serve

On education prioritization:

• “There is no shortcut to just saying ‘we’ve got to start prioritizing education differently’”

On educator roles:

• “Educators are going to socialize, facilitate — what we’re providing is the stimuli”
• Naming and centering the learning goals shifts the focus from what is the mechanism
• Digital scaffolding and the way it supports learners
• “Pedagogical agents” to help make clear what are the consequences of our actions

Teachable agents (Dan Schwartz research area):

• Teaching a virtual agent is a good consolidation of learning
• Having to consult multiple experts that might not always agree

On difficulty calibration:

• Really hard to find the sweet spot that is not dumbed down and not too elevated
• Low match/high match solution was elegant
• How do you convey that the AI is making a prediction against a more trusted source vs. not?

On AI perception and social learning:

• Need to show that AI needs to convert everything to a number — fight growing misconception that AI sees how we see
• AI chatbot companions are “persistent mirrors that change how people see themselves”
• Why do people like AI chat? Whole self across all dimensions?

Closing thought: “What intelligence is and how we learn is the new design problem.”