Event registrations.
Distribution and ranking.
Compare the visible registration counts, from the median event to the largest lists. Each point links to its archived source.
across 1,141 pages
Each dot is one registration page. Horizontal position = visible RSVP counter, on a logarithmic scale. Hover or tap for the event. Snapshot counts are not attendance.
What this ranking can and cannot tell us
These are the latest archived public on-list counters, deduplicated by Partiful page. We have counters for 1,141 pages, not every event. Missing counters are excluded, not treated as zero. A person can appear on many lists. Capacity, invitation rules and capture time differ between events, so this measures visible registrations rather than event quality.
Event topics.
How they overlap.
Which technology conversations overlapped with business problems? Every node is a substantive subject. Every line comes from shared events.
Purple: technology. Blue: business. Nodes require 25 events; links require 8 shared events and at least 1.2 times the overlap expected from label frequencies. Main and secondary subjects only. Happy hour, breakfast and parties are formats. Layout shows association, not growth over time.
Most frequent phrases
in event descriptions.
Phrase size reflects the number of event descriptions containing it. Click a phrase to locate those events.
Topics ranked by number of events
Technology topics
Business themes
Main or secondary subjects, by registration pages after confirmed replacements are merged. One event can appear in several rows. The generic AI label is not the union of all AI subjects.
Event registrations
by company involvement.
A focused comparison of Google, Amazon/AWS, Apple, Meta, Microsoft and NVIDIA. We separate organizer or sponsor roles from speaker affiliations. This is a named-company comparison, not a measure of company size.
One event per group. If a selected company organizes or sponsors an event, it enters that group even when speakers also have a selected-company affiliation. Missing counters are excluded, not treated as zero.
Does the difference remain within a format?
Any role involving the six selected companies versus no recorded link to them. Only formats with at least 10 measured events on both sides are shown. Formats overlap.
Company names, source coverage and limits of the comparison
Topic frequency and
median registrations
Subject frequency on the horizontal axis; median visible RSVP count on the vertical axis. At least 20 measured events per subject. Topics overlap. Tap a subject for coverage and source events.
Topic pairs with the
highest relative overlap
Ranked by lift with at least 10 shared events. Lift compares observed overlap with the overlap expected from label frequencies. Related concepts can naturally have high lift.
Where events
were announced.
See how archived announcements were distributed across mapped areas of San Francisco. Filter by day to compare the geographic pattern.
That is the busiest archived start-hour among timed announcements. Click any hour to inspect the sources. Times are Pacific; occurrence is not independently verified.
Companies as hosts,
sponsors and employers.
Compare organizing and sponsoring with speaker affiliations. A speaker's employer is a separate link, not evidence that the company hosted the event.
Entity and role caveats
A logo alone does not establish sponsorship. Current organizer, sponsor and speaker-affiliation roles are separate from historical, program, judge, mentor and host associations. The all-links sort includes these explicitly labeled roles. Unresolved logo identities are excluded. Organization aliases such as SVB / Silicon Valley Bank are still separate source identities; this is an exploratory comparison, not a final audited company leaderboard.
Companies, events,
speakers and topics.
Companies connect to events through actual roles. Events connect to people and topics. Explore the web, then click a node to see the underlying records.
Explore recorded
connections.
Hover to isolate direct connections.
Click to inspect. Drag to move.
Use + / - to zoom.
Browse graph nodes as a list
People, roles
and social profiles.
Find people by topic, company and industry. See who they are, where they spoke and why they appear in this archive. Build your own shortlist.
Photos, videos
and event recaps on X.
Browse selected public posts about SF Tech Week. Sort by observed views, likes or publication date, and open each original post.
No stories match these filters. Try a broader phrase or choose All stories.
Original posts load as you scroll. Source links always remain available.
Technology demos.
Unusual event formats.
A fly-brain model, agents taking live calls and a manicure with an NFC chip. Twelve specific stories from the program and its public recaps, each with a source.
Six technology demos and experiments
Technical choices you can point to. Reported demonstrations are attributed to their authors.
Six unusual event formats
These formats appeared in the archived listings. An invitation tells us what was planned; it does not prove how the night unfolded.
An editorial selection, not a ranking or evidence of a market trend. Listings were archived October 14; X text was captured October 16, 2026; selected media was retrieved October 17. Technical claims have not been independently reproduced. Read the source notes ↗
Apply AI
in your business.
We help teams turn AI experiments into useful workflows. Bring a process, a bottleneck or an idea. Let's work out a practical first step.
Search the archive from ChatGPT, Claude or Codex. Verify your email to connect. Read-only access, with links to sources.