Transcription
[Music] [Applause] Hi, welcome to another video.
So, Gemini Code Assist has had a steady stream of updates since June 12th. And if you're curious about what's changed, I'm going to walk you through every new feature in order, starting right after June 12th and going all the way to the latest release. There's a lot to cover, especially if you use VS Code or Intelligj. So, let's get into it.
I'll be starting after the June 12th updates because I had already covered everything up to that point in the last video. So, starting with June 18th, Gemini Code Assist chat suggestions are now shown in a preview block by default in VS Code. This makes it much easier to read the code that Gemini generates in chat. And you can even configure the preview pane to show suggestions as collapsed or expanded. Alongside that, you can now revert to a checkpoint for chat code suggestions. So, if you try an AI suggestion and it doesn't work out, you can easily roll your files back to the point before you applied it, which is super helpful for experimenting safely.
On June 25th, there were some big improvements for agent mode in VS Code. You can now use agent mode to tackle complex multi-step tasks with Gemini Code Assist. The agent presents a plan for you to review and approve before making any changes, and you can edit, approve, or deny each suggestion. There's also persistent agent mode and chat state between IDE restarts, so your work doesn't get lost if you close your editor. Multi-file editing is now available in agent mode, letting Gemini update all relevant files in your codebase at once, which is awesome for big refactors or feature implementations. Plus, agent mode now has full project context awareness, meaning Gemini analyzes your entire project for more accurate code completions and suggestions.
Moving on to July 1st, Gemini 2.5 Pro and Gemini 2.5 Flash models became generally available. These are now powering Gemini code assists chat, code generation, and code transformation capabilities. The new models are much better at handling complex code, math, and science tasks. So, you'll notice more accurate and helpful suggestions overall.
On July 17th, some features that were previously in preview became generally available. This includes checkpoints, selected code snippets, terminal output, and multifile editing, so everyone can now use these without having to opt in to a preview channel.
July 18th brought improved code completion speed for VS Code. Gemini Code Assist. suggestions now show up faster, which is always a plus when you're working on something time-sensitive.
On July 24th, Intelligj Gemini Code Assist added the ability to stop in progress chat responses. So, if Gemini starts going off track or you realize you ask the wrong thing, you can halt the response immediately. This update also introduced thinking tokens, which give you insights into Gemini's thought process before it produces a response. So you know it's actively working on your request.
July 28th saw the release of inline diff view in VS Code Gemini code assist. This feature highlights code changes directly in your editor. Green for additions, red for deletions, making it super easy to review modifications and integrate new code without leaving your current view.
On July 31st, agent mode in both VS Code and Intelligj got even better. You can now enable auto approve mode, letting the agent act on your behalf and automatically apply changes. You can review and roll back these changes as needed. Also, agent mode now includes real-time shell command output, improved UI performance, and persistent chat state between IDE restarts.
Moving to August 1st, Intelligj Gemini Code Assist introduced a code diff view directly in chat so you can see additions and deletions highlighted in the conversation. Multi-art chat code suggestions are now supported, letting you accept one change or all at once. Clickable file names in chat output were added, too. So, you can open files directly from the chat. The chat banner now shows your current release channel, which is handy if you're on experimental builds.
On August 8th, VS Code Gemini Code Assist got a quick preview feature for chat code suggestions across multiple files. You can preview all suggested changes in one place and jump straight to the file and suggestion you want, which is a big time for larger projects.
August 12th brought a major upgrade to code customization. You can now set up and manage code customization directly in the Google Cloud Console. Create repository indexes, add repositories, and manage repository groups for granular access control. This is especially useful for teams working across multiple code bases.
Finally, on August 15th, VS Code Gemini Code Assist received numerous performance improvements, including lower CPU and memory usage and less extension slowdown. The VS Code chat banner now shows the release channel so you know which features are available. Intelligj users can now delete individual prompt and response pairs from chat history, making it easier to keep things tidy without losing useful context.
Now, let me show you these features in action as well. But before proceeding, let me tell you about Ninja Chat. Ninja Chat is an all-in-one AI platform where for just $11 per month, you get access to top AI models like GPT40, Claude for Sonnet, and Gemini 2.5 Pro, all in one place. I've been using Gemini for quick research. But what's really cool is their AI playground where you can compare responses from different models side by side. Their mind map generator is a gamecher for organizing complex ideas as well. The basic plan gives you 1,000 messages, 30 images, and five videos monthly with higher tiers available. If you need more, use my code king 25 for 25% off any plan or king 40 yearly for 40% off annual subscriptions. Check the link in description to try it yourself.
Now, back to the video. So, to start, you'll need to update it in VS Code. Just go ahead and do that. Now, first of all, the agent mode is available without messing around with things like insider mode or anything else, which is great. There's also a new checkpoint option. Checkpoints are very similar to what you see in tools like R code or Klein where it makes checkpoints at intervals and you can then revert back if it does something really bad. Another thing is local codebase awareness. Local codebase awareness is something that you can easily enable in the settings and that will basically index your codebase in order to give you better suggestions, be a bit more rapid in finding files and improve overall navigation. It's similar to what you see in R code with codebase indexing as well. Apart from that, you can also now see a quick preview of the snippet of code that it generates or edits accordingly, which is also awesome.
If you use Intelligj, then you'll also now get thinking tokens as well as the ability to stop the generation midway. There's also the inline diff option available, which now allows you to easily see the diff directly in the view. After that, there's also the new auto approve option. Auto approve basically does what it says. It automatically approves all tool calls without asking for any input, which is great for letting it run in the background. You also now get the list of files changed, which you can view in a simple looking modal near the prompt box. That is majorly about it. So, go ahead, use it. and I really like it for a lot of tasks. You can use it all you want without any issues.
Also for MCPs, I only use two MCPS which are context 7 for documentation fetching and biter over for memory. I really like it as a memory layer. It allows you to plug it into all kinds of coders and it can create memories that are synced across all your coders. You can also share it with your teammates, which allows you to build rules and memories that stick to your projects. So, go ahead and use Gemini Code Assist because it's a really good option to use. Overall, it's pretty cool.
Anyway, share your thoughts below and subscribe to the channel. You can also donate via Superthanks option or join the channel as well and get some perks. I'll see you in the next video. Bye. [Music]