scripod.com

Are We Using Opus 5 Wrong?

Overview

Shownote

Highlights

Transcript

Chapters

Pins

Are We Using Opus 5 Wrong?

The Daily AI Show

1 DAYS AGO

Are We Using Opus 5 Wrong?

Are We Using Opus 5 Wrong?

The Daily AI Show

The Daily AI Show

1 DAYS AGO

Overview Shownote Highlights Transcript Chapters Pins

Shownote

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 wei...

Highlights

This episode of the Daily AI Show covers the mixed early reactions to Opus 5, the debate over open-weight AI models, and practical tips for getting better results from newer reasoning models. The hosts also explore the new ChatGPT Sites beta, demonstrating its use for building dashboards, internal tools, and sales prototypes, and discuss the potential of ChatGPT Voice for richer, more natural interactions.

00:00

Opus 5 is praised as the best current model

02:35

Half the price of Opus and a significant leap forward

04:30

Restrictions would harm US open models

06:19

Authoritarian governments could use AI for military superiority or repression

07:31

Risks of powerful models being used for cyber or biological attacks

08:11

Crack down on industrial-scale model distillation

08:34

Anthropic does not support a blanket ban on open-weight models

09:13

Pre-release safety testing is key for cyber, biological, and alignment risks.

12:37

Closing loopholes would help open models

17:09

Old, overly detailed prompts may not work well with Opus 5

19:27

Skills are just markdown files with instructions.

19:42

Adapt prompting skills to unlock Opus 5's potential

20:35

Stripping old skills reduced hallucinations

22:07

Removed something to improve accuracy and clarity

22:46

Building a better version of Salesforce reports

23:35

ChatGPT Sites builds personalized BI dashboards.

27:28

The actual episode count is 777, not 849.

28:23

Search pulls from YouTube data without API validation

30:02

Using Claude to generate a dedicated episode page

32:19

Two paths: one-click publish or 30-minute move to Netlify

34:02

Free hosting locks users in without GitHub backup

35:11

Sensitive content can be unintentionally exposed.

36:20

OpenAI may restrict public access if usage becomes unusually high.

38:08

Most sites get few visits.

39:53

Anthropic fixed the no-index tag and removed exposed chats

41:50

Build a demo during a call using dummy data.

42:54

ChatGPT Sites boost sales close rates.

45:06

Desire for a Jarvis-style AI orchestrator

45:15

ChatGPT Voice acts as a Jarvis-like orchestrator.

47:31

A boom mic would encourage more frequent voice use

48:28

Silent background listening is a key convenience.

50:44

Voice dictation is faster and less mentally taxing than typing

51:30

Advanced models pick up spoken nuances lost in written prompts

52:02

Voice notes capture more nuance than typed drafts.

54:49

I prefer British voices for AI, even though my mother is Australian.

59:25

Functionality matters more than aesthetics

1:01:17

If it works and solves a problem, the process doesn't matter.

1:04:59

Thank you for your support.

Chapters

Episode Intro And Brian Returns

00:00

Opus 5 Early Reaction

02:35

Open Weight AI Alliance

04:24

Dario Amodei Responds To Open Weight Criticism

06:13

Authoritarian Governments And AI Risk

07:31

Chip Restrictions And Smuggling

08:07

Industrial-Scale Model Distillation

08:32

Pre-Release Safety Testing For Powerful Models

09:11

Anthropic, China And Open Model Tensions

12:36

Figuring Out How To Use Opus 5

17:08

Benchmarks Versus Real User Experience

19:06

Old Claude Skills And Overly Restrictive Instructions

19:35

Known Unknowns And Smarter Prompting

20:33

Stripping Claude Skills And Improving Results

22:00

ChatGPT Sites Beta

22:44

Sites For Dashboards And Business Intelligence

23:26

Live Daily AI Show Website Build

27:28

Episode Search And Site Navigation

28:20

Where ChatGPT Sites Gets Its Data

30:00

Site Features, Episode Pages And Publishing

32:18

One-Click Publishing

33:59

GitHub, Portability And Platform Lock-In

35:11

Public AI Sites And Privacy Risks

36:11

Hosting Limits During The Sites Beta

37:59

Shared Claude Chats And Google Indexing

39:53

Publishing The Site Live

41:49

AI-Built Proofs Of Concept For Sales

42:41

Working All Day With ChatGPT Voice

45:01

Voice As A Jarvis-Style AI Orchestrator

45:15

ChatGPT Voice Searches During Conversation

47:29

Microphones And Always-Available Voice AI

48:25

Whisperflow And Voice Dictation

50:41

Voice Uses More Words But Less Mental Effort

51:26

Spoken Prompts Add Context And Nuance

52:00

AI Voice, Accents And Trust

54:44

Moving Sites Through GitHub And Netlify

56:38

Building A CCleaner Replacement With Claude

1:01:11

Website Update And Three-Year Anniversary

1:04:58

Transcript

Brian Maucere: Hey, what's going on, everybody? Welcome to the Daily AI Show. Today is July 28th, 2026. And we're live. It's 10 a.m, if you can tell. 10 a.m. On the East Coast, 7 a.m. On the West Coast. And it's what? Tuesday. Yeah, see, I've been off. So ...