In this episode, Ray breaks down a petition signed by 800+ leaders calling to pause the superintelligence race, then digs into AWS outage chatter about AI replacing DevOps, and Microsoft’s Gaming Copilot privacy concerns. Plus: a prompt trick that boosts LLM accuracy, a 2-billion-FPS camera that visualizes light, and JAXA’s HTV-X cargo launch to the ISS.
-Want to be a Guest on a Podcast or YouTube Channel? Sign up for GuestMatch.Pro
-Thinking of buying a Starlink? Use my link to support the show.
Subscribe to the Newsletter.
Email Ray if you want to get in touch!
Like and Follow Geek News Central’s Facebook Page.

Full Summary
In this episode of the Geek News Central podcast, host Ray Cochrane opens by welcoming listeners to episode 1849. He leads with the new story about over 800 tech leaders, including figures like Steve Wozniak and Richard Branson, signing a petition to pause the development of superintelligent AI due to concerns that its rapid advancement could exceed human control. Cochrane reflects on this unified public statement, highlighting concerns about AI’s societal implications and the need for regulatory oversight, especially given the current inability of infrastructure to support extensive AI supercomputing facilities.
Following this segment, Ray shares personal updates about a busy week at school and some beta testing at Blubrry. He then thanks the sponsor, GoDaddy, and encourages listeners to take advantage of exclusive deals on their website. He reminds listeners about supporting the podcast by signing up for newsletters and becoming insiders.
Transitioning to tech news, Ray discusses an AWS outage that occurred after Amazon ALLEGEDLY replaced a significant portion of its DevOps team with AI, according to internal Slack messages and communications that were “removed”. He conveys skepticism towards the news, stressing the dangers of relying solely on automation without human intervention in critical infrastructure roles. Though it’s currently unconfirmed whether this piece of gossip is genuine, the implications could be concerning, particularly given the idea of large-scale automation of systems that once required significant human oversight.
Ray continues with more news about AI, referencing another article about Microsoft’s gaming Copilot facing privacy backlash after failing privacy audits that revealed it continued to record user data despite being disabled. He further highlights that this new ‘Feature’ is installed without confirmation or request, silently added in the background. This type of invasive behavior continues to highlight privacy issues and the related mass exodus of users from Microsoft OS to Linux this year.
The conversation then moves to research showing improvements in LLM accuracy through a simple prompt that enhances language model responses. This research shows how language models generate outputs based on probability distributions. The prompt can significantly increase the diversity of LLM reactions, according to the study. If you want to know the prompt, tune in!
We then move on to discuss a potential new trend in software development, where a software platform called Woz, a production-ready coding framework, is looking to replace Vibe coding as a successor that actually works. The platform emphasizes creating stable software with human oversight, countering the vagueness often associated with vibe coding.
(I mistakenly thought this product was related to Wozniak, but I guess I just had all my wires crossed since I was talking about him earlier in the show.