Google just made computer use native in Gemini 3.5 Flash — its fastest, cheapest model.
AI agents can now see screens, click buttons, navigate software. No APIs needed.
78.4% OSWorld score. Flash pricing. The $35B RPA industry has a problem.
https://t.co/qgHdFsszT5
Here's your official warning: Do not plan on leaving early on the last day of the conference.
Shelly Sousa is taking the stage to close out Midwest Dreamin' 2026, and if you've seen Shelly speak before, you already know why we're excited. 🙌
Register: https://t.co/Kwm1ZjRFkB
SpaceX's S-1 just exposed something every enterprise AI buyer needs to see.
If you pay Anthropic for Claude or Google for Gemini, you're now an indirect customer of Elon Musk.
Combined monthly compute payments to xAI: $2.17B. Annualized: $26B.
🧵
https://t.co/hcHBnKCHFB
#Podcast🎙️ "AI Precipice: The Pulse Beneath"
A reminder that the most important #AI conversations are rarely the loudest ones.
✅Streaming Tuesday, June 23, 2026
✅Time: 12:00 PM PT / 3:00 PM ET
✅Live Link: https://t.co/zCi5VbsKYH
✅Learn More: https://t.co/zGdPjzQVDo
The 47% vs 9% rule:
Agents shipped without automated evals roll back 47% of the time.
Agents with full eval coverage roll back 9%.
Eval pipelines are now load-bearing infrastructure, not a nice-to-have.
BCG + Bain just dropped 2026 AI agent payback benchmarks.
Median: 5.1 months.
Function-level dispersion is brutal:
• SDR agents: 3.4 mo
• Customer service: 4.1 mo
• Engineering: 9.3 mo
• Legal: 14.8 mo
• Clinical: 18.4 mo
Same model. Different math.
What a way to help kick off the 2026 edition of @MidWest_Dreamin … an opening keynote featuring all of you, in a raw, unfiltered conversation. You REALLLY should be in Minneapolis this July 15-17, or you’ll miss out!