Thanks for listening to the AI Moment podcast today. 

On "The AI Moment" podcast, Jonathan Wagstaffe and I discuss our initial thoughts on the recent release of ChatGPT-5 and other AI models released, such as claude 4.1 and a quick temperature check on Amazon Alexa+.

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 (If you want to change to other models you need the £200/$200 p/m package)

• 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 Jonathan & I 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

General AI Market & Advice:

• Claude has also reduced its prices, following suit with OpenAI. It’s a race to the bottom (already) 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. We emphasise treating AI like a "junior coworker" whose work must be verified for accuracy and truthfulness before use.

• Remember my 3T’s, Time (save you time and give you time back) Truthful (is it truthful in its answer - you should check) and do you Trust it (would you put your name and reputation to the answer or output)

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 ai@dannydenhard.com 

——

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.