AI x Open Media Forum - Refraction Post-Event Report
1. Overview & Context
As a leader in the web3 and culture space, Refraction has a proven track record of producing cultural events, conferences, and art exhibitions worldwide, bringing technology companies and artists together to further expand the web3 ecosystem.
Refraction’s proposal to Livepeer’s Devconnect Assembly RFP was successful, and saw the team produce and co-curate the AI x Open Media Forum, which took place on 18th November 2025 at Valerie’s Factory in Buenos Aires during Devconnect.
With a focus on participatory sessions and lightning talks, the Forum paired creative leaders and technical experts to discuss what’s happening, and what’s next in AI, video, and real-time media.
2. Original RFP Deliverables (Restated)
Refraction’s initial proposal promised:
-
A high quality event of 60-80 participants at Valerie’s Factory
-
An immersive programme focussed on the key themes of Open/Decentralized Social, AI in Media, and Video/Media Infrastructure, discussed in 3 roundtable sessions, complimented by an opening plenary and closing remarks.
-
Facilitation of all aspects of the program presentation, including A/V management, attendee RSVPs and check-in
-
A dynamic repository of event photography and video content for socials
-
A co-published thought leadership piece by Refraction and Livepeer
-
35k USD budget
3. Execution Summary
3.1 Venue & Production
Due to external event calendars, the date of this event was pushed up to Tuesday, 18th November to ensure optimal audience engagement. The AI x Open Media Forum took place at Valerie’s Factory as initially proposed.
In the weeks leading up to the event, Livepeer and Refraction agreed that an on-site A/V team with livestream capacity was necessary to achieve Livepeer’s goals. Although it was difficult to secure the team within the short timeline, Refraction contracted a robust livestream video team, who brought in their own equipment and wifi boosters to ensure a smooth livestream viewing experience.
3.2 Budget
The actual expenditures closely matched the proposed budget, with the addition of the livestream video team. Because this expenditure came in higher than anticipated, but was deemed necessary and reasonable by both parties, the additional cost was split between Refraction and Livepeer.
4. Attendance & Programming Outcomes
Ahead of the event, the RSVP page for the daytime event had 270 responses, with 180 approved to attend. The evening party had 1563 responses, with 938 approved to attend.
On the day of the event, 30 people checked in during the Forum via Luma, plus 20 additional guests and facilitators. Following the Forum, Livepeer was a partner on Refraction’s evening event, at which 273 people checked in
The target attendance range for the daytime Forum event was 60-80 people, and we fell short of that number. The audience that attended were high signal, with several international and local creators, artists, builders and extended Livepeer and Refraction partners and community in attendance.
In our initial proposal, Refraction proposed 3 roundtable sessions, with an opening plenary and closing remarks keynote planned before and after the roundtables. A VIP lunch before the Forum and closing cocktails following the daytime sessions were planned to complement the programme and encourage attendees to network.
As we worked with Livepeer to confirm the sessions, we decided to focus on 2 roundtable sessions, each with their own breakout groups catering to creatives/marketers and developers, respectively.
We also decided to host the keynotes during the opening plenary, and created a Lightning Talks section following the roundtable. Applications to the lightning talks were added to the event RSVP, and Livepeer and Refraction’s teams handpicked the talks to be greenlit.
Our confirmed speakers and session facilitators were Doug Petkanics (Livepeer), Alice Scope (Serpentine Arts Technologies), Malcolm Levy (Refraction), Nick Hollins (Livepeer), Rich O’Grady (Livepeer), Almond Oye (Base), Greg Bresnitz (FWB), Jose Meija (MESH), and Hana Yoosuf (Women in Web 3 Privacy). Due to unforeseen circumstances, Jose Meija could not attend the event.
Our confirmed lightning talk hosts were Kevin Leffew (Coinbase), Dovliueich (Daydream), Atown + Sandy (Emerge), and Erin Magennis (Causality Network).
Ultimately due to lower than expected attendance numbers, the VIP lunch was opened to all attendees, and the 4 planned break out sessions were reduced to 3 to ensure the sessions felt intimate.
5. What Fell Short (Execution Gaps)
Timeline and Promotion
The tight timeline—with the RFP signed off only six weeks before the event—was a significant drawback. This limited the ability to ideate, program, and promote the sessions as desired. More lead time was necessary to secure a marquee keynote and overall event programming. With more time, we could have had a clearer, stronger programme to promote and stimulate more RSVPs and attendance.
Session Attendance and Invite Strategy
Daytime attendance was low, with only 30-35 check-ins against a target of 60 to 80 attendees. The “invite only” strategy, while aiming for a cozy feel, may have been too restrictive for a large conference like Devconnect, and opening up the invites more quickly could have been beneficial. The VIP lunch did not go ahead, as many of the invited guests did not attend.
Facilitator Attendance and Support
One of the planned facilitators dropped out due to unforeseen circumstances, and we were unable to secure a replacement. We should have had a more robust network of facilitators to safeguard the programme. We also could have more strongly reached out to the Livepeer and wider ecosystem development community, as the room was lacking in developer attendance. As such, Nick Hollins, Rich O’Grady, and Doug Petkanics had to step in to support the roundtable conversations.
Lightning Talks Approval Process and Precommitment
The Lightning Talks were highly attended and well-received, and we could have been better prepared to encourage even more participation. Expediting the approval process for the talks in the future could increase uptake. Of the many people who showed interest via Luma, only one person followed through with a lightning talk, highlighting a need for stronger precommitment from facilitators and speakers and reducing risk by having backups.
6. Strategic Outcomes & Value Delivered
The sessions were high signal, impactful, and the intimate format was appreciated by audience members, facilitators, and partners. Attendees were excited to see conversations about creativity, art, and their impact at a builder-focused conference like Devconnect. The format received strong feedback, the live streaming and video capture were executed well, and the evening event was well attended.
While daytime attendance fell below the original target, the Forum succeeded in generating high-quality conversations, relationship-building, and strategic signal that has continued to inform Livepeer Foundation follow-up work post-Devconnect.
Partnerships & Conversations Catalyzed
The Forum placed Livepeer and Daydream in direct conversation with a number of aligned teams working at the intersection of AI, media, infrastructure, and open protocols. Notable conversations initiated or deepened during and immediately following the event included:
-
Coinbase (via x402): open media infrastructure, payments, and distribution surfaces for creator-facing applications
-
Story Protocol: provenance, attribution, and IP frameworks for AI-native media
-
Arweave: long-term storage, provenance, and media archiving for open video
-
XMTP: messaging and distribution layers for live and AI-generated media
-
ENS: identity, attribution, and creator-native primitives
-
Jamit and Emerge: early-stage teams exploring new creator and AI-media applications that align with Livepeer’s infrastructure capabilities
All of these conversations have since progressed into ongoing follow-ups led by the Livepeer Foundation, ranging from exploratory integration discussions to ecosystem collaboration and developer outreach.
Types of Follow-Up Underway
Post-event, the Livepeer Foundation has been actively following up on signals and relationships surfaced during the Forum, including:
-
Continued partner conversations around AI-native video applications, particularly those requiring low-latency inference and real-time media pipelines
-
Ecosystem discussions focused on provenance, authenticity, and attribution for AI-generated and AI-assisted media
-
Builder outreach informed by Forum insights, shaping how Livepeer and Daydream are positioned to developers working across AI, video, and creative tooling
While many of these efforts are still in exploratory phases, the Forum functioned as a valuable catalyst for surfacing aligned actors and sharpening Livepeer’s understanding of where infrastructure-level support is most needed. The event made for a concentrated context in which these projects engaged directly around shared technical and creative constraints relating to Livepeer, accelerated conversations that may otherwise not have materialized in the general Devconnect week.
Strategic Insights for Livepeer’s Ecosystem, Product, and Positioning
Analysis of the roundtable discussions and recordings surfaced several recurring themes that are now informing Livepeer Foundation strategy:
-
Provenance and authenticity emerged as a top-tier concern across both creators and technologists, particularly in the context of AI-generated media and attribution
-
Real-time AI video was repeatedly identified as a compelling frontier, especially for use cases requiring low-latency inference and live interaction
-
Compute access inequality was a recurring concern, particularly among creators outside North America and Europe. Daydream’s accessibility and openness were highlighted as meaningful differentiators
-
Positioning language that resonated most strongly framed Livepeer as “open video infrastructure” and as a network of technologists who actively care about artists and creators, rather than purely technical tooling
These insights have already begun to influence how the Livepeer Foundation communicates Livepeer and Daydream to different audiences, and how it prioritizes ecosystem development and partner conversations going forward.
7. Content & Post-Event Deliverables Status
This table summarizes the status of all content and post-event deliverables explicitly committed to as part of the AI x Open Media Forum RFP, including what has been delivered, what is live with links, and what remains in progress with expected timelines.
-
High-quality in-person event at Valerie’s Factory (Delivered)
-
Immersive programme covering Open/Decentralized Social, AI in Media, and Video / Media Infrastructure (Delivered)
-
Event photography & video media repository (Delivered)
-
Social media campaign (Delivered)
-
Details / Metrics:
-
X: 6,600 impressions · 77 engagements
-
IG: 4,700 impressions · 64 likes · 5 shares
-
Links:
-
Co-published Refraction × Livepeer thought leadership piece (Delivered)
-
Summit presentation video recordings (In progress)
8. Lessons Learned
The AI x Open Media Forum delivered high-signal conversations, strong content capture, and positive qualitative feedback from attendees and partners. At the same time, the execution surfaced several concrete lessons that will directly inform how Refraction approaches future Livepeer collaborations, particularly those tied to large ecosystem events such as Devconnect.
#1: A six-week turnaround from RFP approval to event delivery constrained programming depth, speaker recruitment, and promotional momentum, directly impacting daytime attendance.
In the future, we will aim for minimum 10–12 week lead time for Livepeer ecosystem events to allow for:
-
Clear thematic narrative and session outcomes
-
Securing at least one marquee keynote speaker early
-
More consistent promotion cadence across partners’ channels
-
Lock program structure and speaker slate earlier, enabling stronger outward-facing messaging and clearer value propositions for attendees.
-
Introduce a program readiness milestone (e.g. T–6 weeks) after which promotion shifts from “announcement” to “conversion-focused” outreach.
#2: The invite-only strategy for the daytime Forum was overly restrictive in the context of a large, high-density conference like Devconnect, resulting in attendance falling short of targets.
In the future we will shift to a hybrid invite + open RSVP model, with:
-
Prioritised access for curated guests and ecosystem partners
-
Controlled public RSVPs released earlier to build attendance buffer
-
Define tiered attendance KPIs (e.g. minimum viable, target, stretch) and monitor weekly against RSVPs and confirmations.
-
Implement earlier confirmation nudges (calendar holds, pre-event reminders, soft commitments) to reduce no-shows.
-
Design ancillary elements (e.g. VIP lunches) with attendance contingency plans, ensuring they can flex based on real-time check-ins.
#3: The facilitator pool was vulnerable to last-minute changes, and developer representation in the room was lighter than intended
In the future, we will:
-
Build a deeper facilitator bench for each session, including named backups confirmed in advance.
-
Formalize facilitator recruitment earlier, with clearer expectations around availability and session ownership.
-
Proactively engage Livepeer’s developer and ecosystem communities earlier in the process, using targeted outreach rather than relying primarily on general RSVPs.
#4: Lightning Talks were a highlight, but uptake and follow-through were lower than initial expressions of interest, creating unnecessary risk.
In the future, we will:
-
Shorten and streamline the Lightning Talk approval process to reduce drop-off.
-
Require explicit precommitment from selected speakers (e.g. confirmation deadlines, lightweight speaker agreements).
-
Overbook Lightning Talks slightly and maintain a backup list to mitigate last-minute cancellations.
#5: While the team adapted well on-site, tighter staffing coverage would have reduced reliance on Livepeer team members stepping in to facilitate sessions.
In the future, we will:
-
Assign clear redundancy across critical roles (facilitation, moderation, speaker management).
-
Increase on-site program support staffing for events with participatory formats.
#6: Livestreaming and content capture were highly successful and validated the added investment.
In the future, we will:
-
Treat livestream and content capture as a default requirement, scoped earlier in budgeting and production planning.
-
Align session formats more explicitly with post-event content goals (short-form clips, thought leadership, documentation).
-
Integrate content outputs into defined post-event deliverables with timelines and ownership agreed upfront.
9. Conclusion & Forward Commitments
The AI x Open Media Forum successfully delivered a high-signal, thoughtfully curated convening during Devconnect Buenos Aires, bringing together creative leaders, technologists, and ecosystem partners to explore the future of AI, video, and open media infrastructure. Despite a compressed timeline and lower-than-target daytime attendance, the quality of discourse, participant engagement, and content produced affirmed the relevance and value of the Forum within a builder-focused conference environment.
Refraction met the core objectives outlined in the original RFP, including venue delivery, program curation and facilitation, end-to-end production, and high-quality content capture. The decision to add livestreaming capabilities meaningfully expanded the Forum’s reach and aligned strongly with Livepeer’s mission and goals. The evening event further demonstrated strong community interest and visibility for Livepeer within the broader Devconnect audience.
Remaining Deliverables & Timelines
-
Event recap videos and lighting talk highlights will be posted to the Livepeer YouTube channel in the coming weeks. Refraction will support this videos across all communications channels
-
We will continue to share the co-published thought leadership piece
-
Refraction and Livepeer will ideate on future iterations of the AI x Open Media Forum, bringing in additional partners to extend the capabilities and reach of the event
-
Refraction and Livepeer will continue work on a co-published zine, bringing insights and content capture from the Buenos Aires Forum together in a creative presentation, to future extend the importance and viability of the project
Refraction would like to extend sincere thanks to the Livepeer Foundation, especially Rich O’Grady, Nick Hollins and Doug Petkanics for the trust, collaboration, and shared ambition in delivering this event, as well as to all speakers, facilitators, lightning talk hosts, and attendees who contributed their time, ideas, and energy. The Forum demonstrated the appetite for deeper conversations at the intersection of open infrastructure, creativity, and media, and provided valuable learnings that will directly inform future Livepeer and Refraction collaborations.
