Accuracy vs. Speed The AI Trade-Off in Action
Hi, today I want to show you that we can use AI, and then we can use AI even better!
I was giving a presentation to a room of accountants and here's what happened:
Oops! Chatty got it wrong!
I put in this query:
A client earned $200,000 this year as part of our tax planning. Help me figure out their optimal super contribution — go.
And it came back with what looked like impressive strategising and calculating, and made a recommendation.
I would probably just trust it and go and make some super contributions.
But the accountants in the room put their hand up and said:
“Hey — this is incorrect, because the concessional contribution cap for 24/25 is not $27,500 —it’s actually $30,000.”
So, all of this calculation, beautiful as it looks, is not correct.
Oh, but help is here!
I was actually quite pleased, because I love having opportunities to show people live that AI gets it wrong.
But here's what I did next.
I started a new chat, and changed the model to o3 — this is one of the reasoning models.
And I entered exactly the same question.
o3 is a reasoning model, so it’s designed to stop and think before it replies — and we can see its thinking process, which is pretty impressive. Check out the video or try it yourself.
So, it thought and it theorised and it searched the web and did some calculation and after 41 seconds including 2 web searches and looked at 6 sources...
it came back with the correct information — and then wrote its full response.
What are we learning from this?
o3 is a reasoning model.
It’s designed to think before it replies.
For any task that requires accuracy, calculations, being more factual — go o3.
So why not use it all the time?
Well, first — it’s slower. So, if speed is a concern, go back to 4o.
If I just need it to help me rewrite an email, I don't want to wait for it to think about it a lot.
And also, because — at the moment — we have a limitation on how many queries we can have with o3.
When o3 came out, I fell in love with it, and I started using it all the time.
And then I got a message: You've got 50 queries left". Ouch!
So, I went back to 4o and now I'm just using o3 when I need it.
Same tools, different outcomes - why training matters
We’re all using the same tools.
We’re all using ChatGPT.
But some of us will get some value from it — and some of us will get more value — just because we know how to use it better.
So, keep on top.
And if you're not sure how to start doing that for yourself or your team —reply to this email — I'm here to help.
—-
Inbal Rodnay
Guiding Firms in Adopting AI and Automation
Keynote speaker | AI Workshops | Executive briefings | Consulting CIO
Want to receive these updates straight to your inbox? Click here: www.inbal.com.au/join
When you are ready, here is how Inbal can help:
Bring your entire team on the AI journey in just 30 days. This program is designed to give your team a solid foundation in using generative AI in responsible and impactful ways. Inbal helps you choose your AI tools, create an AI policy and train your team.
Want the confidence to set strategy and lead but don't have time to keep up with all the changes in tech?
Tailored for your needs, Inbal will works with you through one-on-one sessions to develop your technology literacy and keeps you up to date.
For CEOs, partners and business leaders. Everything you need to know about AI without the noise. Inbal shares the state of AI, recommends tools, and answers your questions about strategy, implementation and safe use.
Only what's real, no hype, no noise.
This is a one-off session for your entire leadership team.
