ChatGPT Just Got a Surprise Update, But OpenAI Can’t Explain How It’s Better
Perhaps Sam Altman’s strawberry photo was a secret message after all, because about a week after OpenAI’s CEO posted the photo to social media, the company confirmed that ChatGPT received a surprise update last week that it hadn’t told anyone about.
This was announced on social media by an official ChatGPT account, which revealed that “there is a new GPT-4o model in ChatGPT”, later adding that this is not a new model like a GPT-5 upgrade, but merely an improvement to the existing GPT-4o model.
There is a new GPT-4o model out in ChatGPT since last week. I hope you all enjoy it and check it out if you haven’t already! We think you’ll like it 😃August 12, 2024
In his Model Release Notes We’re given a vague description of what these upgrades entail – namely “Bug fixes and performance improvements” based on “experiment results and qualitative feedback” – which doesn’t amount to much.
OpenAI admits that there is no new capability or specific upgrade it can point to, and even gives itself a reason for not having a quantitative benchmark it can use to explain these smaller AI behavioral tweaks. So we just have to trust it and look at anecdotal reports that GPT-4o now works better than before.
It’s also unclear if this improvement is the “Strawberry” project we’ve been waiting for. That’s a codename that leaked earlier this year for the next ChatGPT model update, and while this latest improvement would count as an update, and perhaps was what Sam Altman was teasing with that strawberry image, leakers were expecting something more substantial for Strawberry.
It does raise the question, however, of how you track AI improvements. You can debate the amount of data an AI has access to, or the processing power of the chipsets powering AI on devices, but what the end user cares about is how useful the AI actually is, which is hard to quantify because it’s based on human perception rather than something you can measure. Let’s hope someone (OpenAI or otherwise) comes up with an effective benchmarking tool soon.
Anyway, if you really want to use the new ChatGPT model, we have good news: it has already replaced the old version of GPT-4o in the tool. So just head over to ChatGPT and you can talk to the new bot just like you did with the last version of GPT-4o – although it has a message limit for free users.
Alternatively, you might want to check out all the Gemini AI upgrades that Google has announced for its upcoming Google Pixel 9 phones. Features like Gemini Live and Add Me already seem really impressive, and we expect the tools to only get better.