Skip to content
Artwork for The Daily AI Show
The Daily AI Show · July 28 · 1 hr 5 min

Are We Using Opus 5 Wrong?

The episode focused on the early reaction to Opus 5, why some users are getting better results than others, and whether older Claude skills and detailed prompts are actually limiting newer reasoning models. The hosts also discussed the debate over open weight AI, Dario Amodei’s response to criticism of Anthropic’s position, chip restrictions, model distillation, and safety testing for powerful models. Much of the second half centered on ChatGPT Sites, including a live website build, publishing, hosting, search, GitHub portability, privacy concerns, and using AI-generated sites for internal tools and sales prototypes. The final discussion covered ChatGPT Voice, voice search, Whisperflow, spoken prompting, and whether talking to AI provides richer context than typing. Key Points Discussed 00:00:18 Episode Intro And Brian Returns 00:02:35 Opus 5 Early Reaction 00:04:24 Open Weight AI Alliance 00:06:13 Dario Amodei Responds To Open Weight Criticism 00:07:31 Authoritarian Governments And AI Risk 00:08:07 Chip Restrictions And Smuggling 00:08:32 Industrial-Scale Model Distillation 00:09:11 Pre-Release Safety Testing For Powerful Models 00:12:36 Anthropic, China And Open Model Tensions 00:17:08 Figuring Out How To Use Opus 5 00:19:06 Benchmarks Versus Real User Experience 00:19:35 Old Claude Skills And Overly Restrictive Instructions 00:20:33 Known Unknowns And Smarter Prompting 00:22:00 Stripping Claude Skills And Improving Results 00:22:44 ChatGPT Sites Beta 00:23:26 Sites For Dashboards And Business Intelligence 00:27:28 Live Daily AI Show Website Build 00:28:20 Episode Search And Site Navigation 00:30:00 Where ChatGPT Sites Gets Its Data 00:32:18 Site Features, Episode Pages And Publishing 00:33:59 One-Click Publishing 00:35:11 GitHub, Portability And Platform Lock-In 00:36:11 Public AI Sites And Privacy Risks 00:37:59 Hosting Limits During The Sites Beta 00:39:53 Shared Claude Chats And Google Indexing 00:41:49 Publishing The Site Live 00:42:41 AI-Built Proofs Of Concept For Sales 00:45:01 Working All Day With ChatGPT Voice 00:45:15 Voice As A Jarvis-Style AI Orchestrator 00:47:29 ChatGPT Voice Searches During Conversation 00:48:25 Microphones And Always-Available Voice AI 00:50:41 Whisperflow And Voice Dictation 00:51:26 Voice Uses More Words But Less Mental Effort 00:52:00 Spoken Prompts Add Context And Nuance 00:54:44 AI Voice, Accents And Trust 00:56:38 Moving Sites Through GitHub And Netlify 01:01:11 Building A CCleaner Replacement With Claude 01:04:58 Website Update And Three-Year Anniversary 01:05:32 Episode Wrap-Up The Daily AI Show Co Hosts: Brian Maucere, Andy Halliday, Beth Lyons, Gareth.

0:00-1:05:31

transcript

No transcript — this publisher did not publish one.

show notes

The episode focused on the early reaction to Opus 5, why some users are getting better results than others, and whether older Claude skills and detailed prompts are actually limiting newer reasoning models. The hosts also discussed the debate over open weight AI, Dario Amodei’s response to criticism of Anthropic’s position, chip restrictions, model distillation, and safety testing for powerful models. Much of the second half centered on ChatGPT Sites, including a live website build, publishing, hosting, search, GitHub portability, privacy concerns, and using AI-generated sites for internal tools and sales prototypes. The final discussion covered ChatGPT Voice, voice search, Whisperflow, spoken prompting, and whether talking to AI provides richer context than typing.


Key Points Discussed


00:00:18 Episode Intro And Brian Returns

00:02:35 Opus 5 Early Reaction

00:04:24 Open Weight AI Alliance

00:06:13 Dario Amodei Responds To Open Weight Criticism

00:07:31 Authoritarian Governments And AI Risk

00:08:07 Chip Restrictions And Smuggling

00:08:32 Industrial-Scale Model Distillation

00:09:11 Pre-Release Safety Testing For Powerful Models

00:12:36 Anthropic, China And Open Model Tensions

00:17:08 Figuring Out How To Use Opus 5

00:19:06 Benchmarks Versus Real User Experience

00:19:35 Old Claude Skills And Overly Restrictive Instructions

00:20:33 Known Unknowns And Smarter Prompting

00:22:00 Stripping Claude Skills And Improving Results

00:22:44 ChatGPT Sites Beta

00:23:26 Sites For Dashboards And Business Intelligence

00:27:28 Live Daily AI Show Website Build

00:28:20 Episode Search And Site Navigation

00:30:00 Where ChatGPT Sites Gets Its Data

00:32:18 Site Features, Episode Pages And Publishing

00:33:59 One-Click Publishing

00:35:11 GitHub, Portability And Platform Lock-In

00:36:11 Public AI Sites And Privacy Risks

00:37:59 Hosting Limits During The Sites Beta

00:39:53 Shared Claude Chats And Google Indexing

00:41:49 Publishing The Site Live

00:42:41 AI-Built Proofs Of Concept For Sales

00:45:01 Working All Day With ChatGPT Voice

00:45:15 Voice As A Jarvis-Style AI Orchestrator

00:47:29 ChatGPT Voice Searches During Conversation

00:48:25 Microphones And Always-Available Voice AI

00:50:41 Whisperflow And Voice Dictation

00:51:26 Voice Uses More Words But Less Mental Effort

00:52:00 Spoken Prompts Add Context And Nuance

00:54:44 AI Voice, Accents And Trust

00:56:38 Moving Sites Through GitHub And Netlify

01:01:11 Building A CCleaner Replacement With Claude

01:04:58 Website Update And Three-Year Anniversary

01:05:32 Episode Wrap-Up


The Daily AI Show Co Hosts: Brian Maucere, Andy Halliday, Beth Lyons, Gareth.