HeyJaredWorkAITV
A weekly 22–25 minute show on enterprise AI, filmed at the New York Stock Exchange and built for the senior executives.
- Indexed videos, last 90 days
- 25
- Latest publication
- Sep 24, 2026
- Audience
- ~1.2K subscribers
- Earliest in this view
- Aug 20, 2026
Latest videos
Four AI futures marketers should prepare for (opens the original)
Read excerpt
AI shopping could develop around open marketplaces, retailer-controlled platforms, a dominant super app or renewed trust in human creators. Each would change how brands get discovered and how customers decide what to buy. David Bratslavsky walks through the four scenarios and the preparation they share: product information that AI tools can read and a clear reason for people to choose the brand. The report features BCG's scenario framework and OpenAI's shopping demonstrations. Companies featured
Sovereign AI and the captive startup model (opens the original)
Read excerpt
A company that helps build an AI startup may want to own the technology that gives it an advantage. Vantora's sovereign AI model gives enterprise partners an equity stake from the start and a path to bring the business in-house once it proves its value. Using J.B. Hunt as the example, David Bratslavsky explains how private data, an embedded engineering team and an initial customer relationship fit together. The aim is to give the partner more control over AI built around its own operations. Comp
How Vantora builds AI ventures (opens the original)
Read excerpt
Vantora starts with an expensive problem inside a large company, then builds an AI business to solve it. After raising more than $100 million, the company is expanding an approach that puts builders alongside enterprise teams and their operating data. David Bratslavsky looks at the model through the example of Alaska Airlines, how Vantora uses its COSMOS technology and the financial opportunities it says it targets. Vantora reports that it has launched 17 ventures. Company featured: @AlaskaAirli
How embedded evaluators test AI (opens the original)
Read excerpt
Independent safety testers are working inside AI labs while models are still being developed. Accenture and Anthropic are building a team of embedded evaluators to test how models respond to deceptive or harmful requests. David Bratslavsky explains how red teaming works, how testers report failures to engineers and why testing continues after a fix. Passing an individual test does not establish that a model is safe; the work aims to uncover problems before release. Companies featured: @Accenture
TabPFN 3.5 turns business data into predictions (opens the original)
Read excerpt
Supplier records and payment histories can help a business forecast risk, but preparing spreadsheet data for AI often takes substantial work. TabPFN 3.5 is designed to make predictions from rows and columns without a new round of model training for every task. David Bratslavsky looks at SAP's approach to missing entries and mixed formats, plus how businesses can access the Plus version through SAP AI Core. The report covers potential uses including supplier assessments and late-payment forecasts
Publishing over time
Last 90 days. Choose a month to open its work.
Recurring subjects
Named in the text we hold. One piece can cover several.
Audience
~1.2K subscribers
Measured Sep 19, 2026
Source's subscribers, not the number who saw an individual piece.
About this data
Counts cover the work we have indexed. Tone needs enough text and a confident classification. Excerpts and episode notes are not full articles or transcripts.
Identity or attribution wrong? Suggest a correction.