Sign up to save your podcasts
Or
Share ChatGPT5 First Thoughts
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
ChatGPT5 First Thoughts
ChatGPT-5 Key Points:
Simplification: OpenAI has merged its various models into a single ChatGPT-5, simplifying the user interface. An auto-selector feature now analyses the prompt and chooses the appropriate processing model, leading to greater use of its "deep thinking" capabilities.
Legacy Access: After initial backlash, OpenAI has allowed users to revert to older models like GPT-4.
Performance: The new model feels quicker, and token usage is cheaper.
Expectations vs. Reality: The update is seen as an incremental improvement rather than the "game-changing step change" many anticipated.
- Accuracy: OpenAI claims it hallucinates less, but Danny & Jonathan both stress the continued need for human verification. Getting accurate results requires more precise context in prompts. The model is also reportedly less overly agreeable than previous versions. Remember the 3Ts - time you need to invest into the output, truth - how truthful is the answer and trust - do you trust the output enough to put your own name and reputation on it
Tools discussed
- www.chatgptm.com
- www.claude.ai
- Amazon Alexa AI
General AI Market & Advice:
Claude has also reduced its prices, following suit with OpenAI. It’s a race to the bottom in terms of pricing
The hosts note that the AI landscape is a constant "arms race," meaning the "best" model changes frequently and depends on the specific task.
The main advice for users is to experiment with different platforms (ChatGPT, Claude, etc.) to find the one LLM you prefer. They emphasise treating AI like a "junior coworker" whose work must be verified for accuracy and truthfulness before use.
Want to connect with Danny Denhard & Jonathan Wagstaffe
Danny on LinkedIn - https://www.linkedin.com/in/dannydenhard/
Jonathan on LinkedIn - https://www.linkedin.com/in/wagstaffe/
Do you have feedback or questions email us [email protected]
Remember to hit follow or subscribe in your favorite podcast app and if you would like to help spread the word, please share the AI Moment podcast or rate and review the podcast.
View all episodes
By Danny DenhardChatGPT-5 Key Points:
Simplification: OpenAI has merged its various models into a single ChatGPT-5, simplifying the user interface. An auto-selector feature now analyses the prompt and chooses the appropriate processing model, leading to greater use of its "deep thinking" capabilities.
Legacy Access: After initial backlash, OpenAI has allowed users to revert to older models like GPT-4.
Performance: The new model feels quicker, and token usage is cheaper.
Expectations vs. Reality: The update is seen as an incremental improvement rather than the "game-changing step change" many anticipated.
- Accuracy: OpenAI claims it hallucinates less, but Danny & Jonathan both stress the continued need for human verification. Getting accurate results requires more precise context in prompts. The model is also reportedly less overly agreeable than previous versions. Remember the 3Ts - time you need to invest into the output, truth - how truthful is the answer and trust - do you trust the output enough to put your own name and reputation on it
Tools discussed
- www.chatgptm.com
- www.claude.ai
- Amazon Alexa AI
General AI Market & Advice:
Claude has also reduced its prices, following suit with OpenAI. It’s a race to the bottom in terms of pricing
The hosts note that the AI landscape is a constant "arms race," meaning the "best" model changes frequently and depends on the specific task.
The main advice for users is to experiment with different platforms (ChatGPT, Claude, etc.) to find the one LLM you prefer. They emphasise treating AI like a "junior coworker" whose work must be verified for accuracy and truthfulness before use.
Want to connect with Danny Denhard & Jonathan Wagstaffe
Danny on LinkedIn - https://www.linkedin.com/in/dannydenhard/
Jonathan on LinkedIn - https://www.linkedin.com/in/wagstaffe/
Do you have feedback or questions email us [email protected]
Remember to hit follow or subscribe in your favorite podcast app and if you would like to help spread the word, please share the AI Moment podcast or rate and review the podcast.