Uploaded February 2026 | Updated September 2026, 3 weeks ago

---

Stephen Wolfram hosts live and unscripted Ask Me Anything about business, innovation and managing life. Find the playlist of Q&A's here: https://wolfr.am/youtube-sw-business-qa

Originally livestreamed at: twitch.tv/stephen_wolfram

If you missed the original livestream of this episode, feel free to submit a question you would like Stephen to answer in a future Q&A livestream here: https://wolfr.am/12cczmv5J

00:00 Start stream
1:18 SW starts talking
1:33 What do you think about this business idea:
"Is This Worth a Meeting?" App
Upload agenda and get a probability score that the meeting could have been an email.
12:12 What do you think about this business idea:
"AI Co-Founder That Argues With You"
Designed to disagree, poke holes, and slow you down when you're overconfident.
16:02 Could there be a viable business in training AI systems to refuse low-quality questions and help users reformulate better ones?
19:00 How do you know whether AI is actually creating value or just sounding impressive?
25:25 Would you say prompting more like programming, teaching, or having a conversation?
27:12 Do good prompts matter, or does the model mostly figure things out anyway?
27:50 Is there a chance AI companies go bankrupt?
35:30 What's the best way to "test" what an AI is good or bad at?
36:30 Does giving AI a personality help or just confuse things?
38:44 How do you personally use LLMs (local or services) in your day to day work? And for what kind of tasks?
42:31 I've been pretty surprised we haven't seen more FPGA accelerators for cloud database computations
48:41 Are we learning better communication skills (or worse ones) by using AI?
51:42 How protective can we make our IP if we use publicly available models?
54:34 Business idea: Human-in-the-Loop Consulting. Help companies decide where humans should stay in the loop as AI expands.
57:44 What do you think of the recent Claude Opus release? It seems they have improved a lot. If it is just a word predictor, how it would also be able to solve the logical problems? Just fascinated.
1:00:08 Have you considered some benchmarks involving ruliology? Maybe seeing what an embedded agent can discover within its system?
1:03:36 Thoughts on integrating the Wolfram Knowledge Engine into LLM Architecture to cut down on multiple queries to outside databases?
1:05:30 What will happen to a company when their LLM says that it is conscious?
1:08:51 Is it conceivable that LLMs could themselves develop "communication protocols / interfaces" that make sense only to them?
1:13:43 ​​human language are so different and can express so many different things , since llms are mostly running on a english "mindset" what are our blindspots which we would see if we would include more languages
1:18:21 Do you think that llms could enhance linkers or compilers?
1:21:24 End stream

Follow us on our official social media channels.

X: https://x.com/WolframResearch/
Facebook: facebook.com/wolframresearch
Instagram: instagram.com/wolframresearch
LinkedIn: linkedin.com/company/wolfram-research
Stephen Wolfram's Twitter: twitter.com/stephen_wolfram

Contribute to the official Wolfram Community: community.wolfram.com
Stay up-to-date on the latest interest at Wolfram Research through our blog: blog.wolfram.com
Follow Stephen Wolfram's life, interests, and what makes him tick on his blog: writings.stephenwolfram.com