OpenAI Quietly Released a Better ChatGPT Version Surprising Users

[Music] if you've been feeling like your AI buddy's been acting a bit different lately maybe quicker sharper and just a tad smarter you're not alone open AI has been sneaky rolling out some major changes without a big announcement but don't worry I've got all the details you need to know right here so let's talk about it last week I started noticing that chat GPT felt different it was like the responses were more on point faster and just generally better I wasn't the only one either people all over social media were talking about how chat GPT

seemed to be upgraded but here's the thing open aai didn't say a word about it at first it was all very hush hush until they finally dropped a little bombshell on US Open AI took to X to casually mention that they'd slipped in a new version of their GPT 40 model into chat GPT so they just updated the model we've all been using without making a big deal about it the message was simple there's a new GPT 40 model out in chat GPT since last week hope you all are enjoying it and check it out

if you haven't we think you'll like it that's it no fancy press release No Grand unveiling just a tweet typical open AI right now if you're wondering what's so special about this new model let's break it down the updated version of GPT 40 which they're calling chat GPT 40 latest is essentially a fine-tuned and optimized version of what we had before but here's where it gets interesting while open AI hasn't spilled all the beans there's a lot of speculation about what this new model actually is some people out there are thinking this might be part

of a bigger strategy by open AI to release different sized models kind of like what Google and anthropic are doing there's been talk about a GPT 40ar and some think this latest update could be a stepping stone in that direction but I'm not totally sold on that idea because let's be real if it were a brand new model they probably would have hyped it up a lot more so what can this new model do well from what I've seen and what others have reported it's performing better on tasks that require complex reasoning and creativity like

if you've been asking chat GPT to help with coding or solve tricky problems you might have noticed it's just a little bit sharper now it's also faster which is a nice bonus but of course it's not perfect there are still some weird quirks for example in one test the model was asked to stack a book nine eggs a laptop a bottle and a nail in a stable manner the solution it suggested putting nine eggs on top of a bottle I mean come on who does that and then when it was asked how many RS are

in the word strawberry it came back with two which is definitely wrong so yeah there are still some bugs to work out but overall the update is a step in the right direction now talking about strawberry let's talk about something that's been generating a lot of hype project strawberry the idea behind project strawberry is that it could be a new posttraining method that boosts the model's reasoning skills some people are even saying that the improvements we're seeing in chat GPT might be the first signs of this mysterious ious project in action one of the coolest

things about the new chat GPT 4 o latest model is how it handles multi-step reasoning this basically means the AI isn't just jumping to conclusions it's thinking things through step by step before it gives you an answer that's a pretty big deal because it leads to more accurate and thoughtful responses which is something we all want right the new model has already made waves in the AI Community especially in something called the LMS y leaderboard now if you're not familiar with it the LMS leaderboard is like the Olympics for AI models they put different models

head-to-head in all sorts of tasks and the new chat gp40 latest model just crushed it it scored a whopping 1314 points which is the highest score ever recorded on that leaderboard this means it's outperforming some of the biggest names in the game like Google anthropic and meta now if you're thinking how do I get my hands on this new model well it's super easy open AI has already swapped out the old GPT 40 with the new version in both the chat GPT website and app so all you have to do is fire up chat GPT

and you're good to go if you're on the free plan you might hit some message limits but for those of you who are on the plus plan you can push the model to the Limit and really see what it can do but don't worry if you're not ready to Shell out the $20 a month for the plus plan you can still get a good feel for the new model before you hit those limits and then if you run out of messages you can switch over to GPT 40 mini it's not quite the same but it's

still pretty powerful also one more really interesting thing is how open AI has been testing these updates they've been sneaking experimental models into places like the LMS y's chatbot Arena under random names so people don't even realize they're testing new tech the chat GPT 4 o latest model for example was tested under the name Anonymous chatbot and it got over 11,000 votes from users that's a lot of people unknowingly helping out with the testing which just goes to show how clever open ai's approach is so what's next well if this update is anything to go

by we can expect open AI to keep refining and improving chat GPT they're clearly focused on making it better at reasoning creativity and all those tasks that require a bit more brain power and who knows maybe we'll see even more of project strawberry in the future all right now I also want to talk about a new AI model that just came out but it didn't really get the attention it deserves this model called Falcon Mamba 7B was released by The Technology Innovation Institute tii in Abu Dhabi tii is known for working on cutting Edge Technologies

like AI Quantum Computing and Robotics and now they've dropped this new model it's available on hugging face and it's an open-source model which is pretty cool but what really sets it apart is the new architecture it's using most of us are familiar with Transformer models which have been dominating the AI scene for a while now but Falcon Mamba 7B uses something different called the Mamba State space language model slm architecture this new approach is quickly becoming a solid alternative to those traditional Transformer models now why is this important well Transformers are great but they have

some issues especially when it comes to handling longer pieces of text you see Transformers use an attention mechanism that looks at every word in a text and Compares it to every other word to understand the context but as the text gets longer this process demands more and more computing power and memory if you don't have the resources to keep up the model slows down and struggles with longer texts this is where sslm comes in unlike Transformers sslm doesn't just rely on comparing words to each other instead it continuously updates a state as it processes the

text this means it can handle much longer sequences of text without needing a ton of extra memory or computing power now Falcon Mamba 7B uses this sslm architecture which was originally developed by researchers at Carnegie melon and Princeton universities what's cool about this model is that it can dynamically adjust its parameters based on the input so it knows when to focus on certain parts of the text and when to ignore others so how does Falcon Mamba 7B stack up against the big players like meta llama 38b llama 318b and Mistral 7B tii ran some tests

and the results are pretty impressive in terms of how much text the model can handle Falcon Mamba 7B can fit larger sequences than the Transformer models using just a single 24gb A10 GPU this means it can theoretically handle infinite context length if if you process the text token by token or in chunks and again Falcon Mamba 7B came out on top it beat mistol 7bs sliding window attention architecture by generating all tokens at a constant speed without any increase in memory usage and that's a big deal for anyone working with large scale AI tasks because

it means the model is both fast and efficient even when it comes to standard industry benchmarks Falcon Mamba 7B holds its own in tests like Arc truthful QA and GSM 8K it outperformed or matched the top trans former models sure there were a couple of benchmarks like mlu and H swag where it didn't quite Take the Lead but it was still right up there with the best of them but here's the thing this is just the beginning for Falcon Mamba 7B tiu has big plans to keep optimizing the model and expanding its capabilities they're not

just stopping at sslm they're also pushing the limits of Transformer models to keep driving innovation in AI so if you're into AI or just curious about what the future holds keep an eye on falcon 7B it's already making a name for itself and with TI's continued efforts it's only going to get better plus with over 45 million downloads of their Falcon models tii is proving that they're a major player in the AI world all right if you found this interesting make sure to hit that like button subscribe and stay tuned for more AI insights thanks

for watching and I'll catch you in the next one