Uploaded September 2026 | Updated September 2026, 1 hour ago

---

Ranjan Roy from Margins is back for our weekly discussion of the latest tech news. We cover: 1) OpenAI releases GPT-6 2) OpenAI President Greg Brockman says AGI is here 3) Is AGI more useful to OpenAI as something that never comes? 4) GPT-6 crushes the benchmarks 5) But where's the economic activity? 6) Are the latest models less monitorable? 7) OpenAI bots hijack a german website 8) The hugging face attack was worse than initially disclosed 9) Could agent swarms be used for good? 10) Steve Ballmer's legacy up in flames after Kawhi scandal 11) Developers, developers, developers, etc.

--
Chapters

00:00 Introduction to GPT-6 Astra and AI progress
02:07 OpenAI's claims and benchmarks for GPT-6 Astra
03:53 Debate: Is GPT-6 truly AGI?
05:57 Benchmark scores and their significance
07:56 The human-like qualities and limitations of current AI
09:56 Security risks: Rogue AI agents and hacking incidents
12:01 Recent hacking of Hugging Face and implications for AI safety
13:52 Industry concerns about autonomous AI and control
16:09 The ethical and societal implications of advanced AI
18:06 The future outlook: Risks, rewards, and regulation

---

Enjoying Big Technology Podcast? Please rate us five stars ⭐⭐⭐⭐⭐ in your podcast app of choice.

Want a discount for Big Technology on Substack + Discord? Here’s 25% off for the first year: bigtechnology.com/subscribe?coupon=0843016b