024: OpenAI Drops Open Source Models, Grok Makes NSFW Deepfakes, AI-Proof Jobs
Manage episode 498780893 series 3676774
From AI with Kyle News and Updates Live Stream.
First aired: 6 August 2025
Watch full edited live stream: https://youtu.be/3djXXQuDQYw
Get the full notes and summary: https://newsletter.aiwithkyle.com/subscribe
Subscribe and tun on notifications to catch the next live stream: https://www.youtube.com/@iamkylebalmer?sub_confirmation=1
Main Topics & News
0:00 - Introduction & Weekly AI News Overview
1:00 - Ben's Bites Newsletter & AI News Roundup
2:00 - GPT-5 Prediction for Thursday Release
3:00 - OpenAI's New Open Source Models Announcement
5:00 - GPT-OSS Models Deep Dive & Analysis
7:00 - Safety Considerations for Open Source AI
9:00 - Trump's AI Action Plan & US vs China Competition
11:00 - Why GPT-OSS is More Important Than GPT-5
12:00 - Q&A: Phone Performance & Technical Discussion
14:00 - Host Background & Business Focus Explanation
16:00 - GPT-OSS Model Details & Downloads
18:00 - Local AI Performance Issues & Server Solutions
20:00 - Solopreneur AI App Development Advice
22:00 - Claude Opus 4.1 Release Analysis
25:00 - AI Company Market Share & Google Trends
28:00 - Grok's New "Imagine" Feature & NSFW Content
32:00 - UK MP Creates AI Avatar Chatbot
36:00 - AI-Proof Career Advice & Job Market Analysis
42:00 - Adult Content AI Industry Discussion
46:00 - Local vs Online AI Model Security
49:00 - Open Source AI Modification & Licensing
51:00 - ChatGPT Privacy Settings & Data Training
53:00 - Manus vs ChatGPT Agent Comparison
55:00 - Open Source Model Memory Limitations
56:00 - JSON Prompting for Image Generation
59:00 - Closing & Tomorrow's GPT-5 Expectations
Key Questions Answered
12:00 - "Does the 20B model work on phones?"
14:00 - "This guy's not technical, he did an arts degree"
16:00 - "What does GPT-OSS stand for?"
20:00 - "Solopreneur working on mortgage app - advice?"
32:00 - "You still get flagged as an AI avatar?"
36:00 - "What AI jobs involve no coding experience in the UK?"
42:00 - "Thoughts on AI gooning?"
45:00 - "Is running offline LLMs more trustworthy?"
48:00 - "Can you modify open source models?"
50:00 - "If we switch off memory in ChatGPT, does it still train?"
52:00 - "What's your opinion about Manus?"
54:00 - "Do open source GPT models have user memory?"
56:00 - "Is JSON description more precise for image generation?"
21 episodes