A deep dive into GPT 4.5 model release, its performance, industry reactions, and the future of scaling large language models.
Check out a clip from yesterday’s bonus episode: https://youtu.be/AiRUVgTBHHg
Svicpodcast.com: For Questions, Book recs, Videos, Newsletters, etc.
patreon.com/svic
Follow us:
Threads: www.threads.net/@svicpodcast
Twitter: https://x.com/svicpodcast
LinkedIn: https://www.linkedin.com/company/svicpodcast
Tiktok: https://www.tiktok.com/@svicpodcast
Instagram: https://www.instagram.com/svicpodcast
Spotify: https://podcasters.spotify.com/pod/show/svicpodcast
Show hosts Joe (Eng VP) and Jordan (M&A Deal Lead) worked at various companies from Google, Apple, Facebook, Microsoft, Salesforce, Slack, Carta, Splunk, Wealthfront, Adobe, and more.
#ai #openai
Chapters
00:00 – Initial reactions to the new AI model
03:01 – OpenAI’s Live Stream Announcing GPT-4.5
14:25 – Evaluating Grok 3’s performance and industry positioning
16:31 – Concerns over the new model being labeled a “preview”
17:17 – Are we hitting the limits of pre-training performance gains?
18:51 – The cost vs. benefit of scaling up AI models
19:28 – System cards and lack of meaningful disclosures
20:12 – Did XAI find a more effective training approach?
21:44 – Sam Altman’s views on the future of AI improvements
23:11 – Speculation vs. reality in AI advancements
24:05 – The shifting narratives around AI risks and AGI timelines
25:32 – Media hype and misinformation in AI discussions
27:37 – Reviewing the “P-Doom” scores of AI leaders
30:51 – Should AI doomsayers be held accountable for their claims?
32:19 – The cost of bad hires and Zappos’ reverse signing bonus
33:19 – Microsoft, SoftBank, and OpenAI’s shifting alliances
35:48 – Balancing AI investment, overbuilding, and real-world adoption
36:28 – GPT-4.5 pricing and OpenAI’s evolving strategy
37:16 – Training optimizations and reducing AI model costs
39:32 – The future of AI architectures and new research directions
40:38 – Final thoughts and closing remarks
