5 things Gemini still can’t do that keep me paying for ChatGPT

Gemini is the Swiss Army knife of the AI world. Backed by Google, it is available across platforms and deeply integrated into Android, letting it do things no other chatbot can do right now. And the pace at which Google has brought Gemini up to speed with its rivals is just remarkable. Having said that, it often feels like ChatGPT made the best of its first-mover advantage, maintaining its lead in overall usability and response quality even today.

I jump between Gemini and ChatGPT throughout the day for different tasks, using each quite extensively. Despite having access to Gemini’s full capabilities through my Google AI Pro plan, I continue to separately subscribe to ChatGPT because Google’s AI chatbot still doesn’t feel like it’s there yet, and these are my reasons.

There have been several instances where I would give Gemini a fairly lengthy set of instructions, and it would just skip a few of them as if they were never there. That has happened within individual chats and even inside Notebooks that are supposed to pull together context from multiple chats. I experimented with running a full-fledged expense manager inside Gemini recently, and I faced this issue more than once where it would just refuse to follow instructions.

Additionally, Gemini’s scheduled tasks feature has given up on me entirely for the past couple of weeks. It just wouldn’t run despite me giving commands to restart the schedule on multiple occasions and even turning on the toggle manually. Whether it is a bug or something else, I’m not sure, but the fact is that Gemini feels like it’s still a work in progress and could use some polishing before actually feeling reliable.

I would have forgiven it easily for such relatively minor hiccups, but I’ve had a far more consistent and reliable experience with ChatGPT over the last few years to put up with whatever’s going wrong with Gemini in the backend.

This one is a bit subjective, but it does matter a lot to me since I’ve been using ChatGPT for what feels like forever. It has a lot of my personal context that I have intentionally shared with it through hundreds of personal and work chats.

This is something very hard to replicate when starting with a new chatbot. And Gemini, despite its extensive Personal Intelligence capabilities through the broader Google ecosystem, hasn’t been able to do it for me as of now. Gemini does have a built-in tool to help you import your chats and personal context from other AI chatbots, but I’ve never felt compelled enough to completely switch over, given the reliability issues I consistently face.

Armed with Nano Banana, Gemini has been able to create and edit some remarkable imagery. However, once you have tried ChatGPT’s newly released image generation model, you’ll find the Google alternative lagging quite a bit behind. When editing photos or generating vectors, I have found 10 out of 10 times that Gemini messes up the colors, even if I have given it specific hex codes. And it makes photos look either too dull or unnecessarily bright after a simple touch-up job.

While I am not saying that ChatGPT does a flawless job like a professional photo editor would, it surely feels better at adhering to my instructions and keeping the color palette consistent across edit cycles side by side. Gemini does feel a lot faster at generating images, but I don’t mind waiting a minute extra for ChatGPT’s visibly better output.

Being a Google product, I automatically assumed that Gemini would be the best tool for web search, given how much internet data Google Search already has, but that is surprisingly not true. In day-to-day life, I rely on AI tools to get my morning news fix, including a rundown of all the tech news from the previous day.

Right when I start my work, both Gemini and ChatGPT are tasked with looking up a certain set of websites that I closely follow and giving me a quick summary of everything important to glance over. Even after my repeated prompts to put together a more detailed list that doesn’t skip even the most minor Android tweaks, Gemini continues to miss several important pieces of information that ChatGPT more consistently picks up. It’s a shame how lacking Gemini feels when it comes to pulling up information from the internet in real time.

Gemini is great if you live within the Google ecosystem. It can connect with Gmail, Docs, and Keep to help with your productivity needs, and even Google Flights and Hotels to help plan a trip on a budget. Google even has its own video generation and editing tools. However, it has yet to establish a meaningful connection with the external world, which is critical when many of us rely on several non-Google apps for everyday tasks, something ChatGPT gets right.

For instance, while Gemini can only help you with music discovery and playlist generation through YouTube Music, ChatGPT can do that on both Spotify and Apple Music. It connects with a bunch of third-party travel services like Expedia and Skyscanner, while also letting you use DoorDash and Uber from within the app. I am particularly a fan of its connection with Canva (which is my go-to image editing tool), where I can simply prompt ChatGPT to generate and edit stuff on my behalf.

Having used both of them quite a lot, I’ve realized that I often consciously push myself to use Gemini for certain tasks. Whereas ChatGPT is just instinctive. It’s so baked into my muscle memory that if an AI-worthy question pops into my head, chatgpt.com is the URL my fingers automatically type into the address bar. That says a lot about the kind of habit shift Gemini will have to bring about, and I trust that it’s not going to be an easy feat.