Hello, happy Thursday, and welcome to Issue #9 of VibeCoding.sh. I’ve made an addition to this issue that I think everyone will probably dig, it’s the first section you’ll see below, aptly named, “happenings in the AI world that you should know about.”

With so many announcements in the AI world, now literally by the day, I thought I’d kick things off in each issue with the top news stories that I think everyone should know about, makes sense right?

Also, not to rub it in, but if you’re not a Vibe++ subscriber you are really missing out this time around. I just got finished deleting a ton of content that was included in the Vibe++ edition of this issue. For the price of a latte you can upgrade your subscription and get a lot more content with a focus on actionable tips and tricks you can use to really take your vibe coding skills to the next level.

Of course there will always be a free version of my newsletter but the Vibe++ subscribers really did hit the jackpot this week with a LOT more content than free subscribers, just sayin’ 🤷‍♂️

Okay, but you’re here, you’re hungry for some AI news and vibe coding content, so let’s get to it - thanks for reading and enjoy!

Happenings in the AI world that you should know about

A lot happened in the AI world over the last two weeks, here’s some things that you should definitely know about as they can, and likely will impact your vibe coding workflow now or in the very near future.

OpenAI released o3 and o4-mini

I started using o3 shortly after it was announced and holy moly, it’s good, like crazy ridiculously good. And yes, as you probably guessed, companies like Windsurf are all over this, even offering o4-mini for free immediately.

As for the TL;DR on the new models - here you go, fresh from OpenAI’s announcement post from earlier today ⬇️

OpenAI o3 is our most powerful reasoning model that pushes the frontier across coding, math, science, visual perception, and more. It sets a new SOTA on benchmarks including Codeforces, SWE-bench (without building a custom model-specific scaffold), and MMMU. It’s ideal for complex queries requiring multi-faceted analysis and whose answers may not be immediately obvious. It performs especially strongly at visual tasks like analyzing images, charts, and graphics. In evaluations by external experts, o3 makes 20 percent fewer major errors than OpenAI o1 on difficult, real-world tasks—especially excelling in areas like programming, business/consulting, and creative ideation. Early testers highlighted its analytical rigor as a thought partner and emphasized its ability to generate and critically evaluate novel hypotheses—particularly within biology, math, and engineering contexts.

OpenAI o4-mini is a smaller model optimized for fast, cost-efficient reasoning—it achieves remarkable performance for its size and cost, particularly in math, coding, and visual tasks. It is the best-performing benchmarked model on AIME 2024 and 2025. In expert evaluations, it also outperforms its predecessor, o3‑mini, on non-STEM tasks as well as domains like data science. Thanks to its efficiency, o4-mini supports significantly higher usage limits than o3, making it a strong high-volume, high-throughput option for questions that benefit from reasoning.

Google announced a ton of stuff at Google Cloud Next

Arsh does a good job breaking down the top announcements in this tweet, I’d give it a read.

OpenAI released ChatGPT 4.1

Yes, the naming convention ChatGPT uses for it’s models makes even less sense as time goes on. Their most recent release, 4.5, is now followed by a new, even more powerful model, 4.1 - make it make sense 🫠

While the naming convention makes no sense, the model itself is insanely impressive and vibe coders are quickly switching to it in their favorite vibe coding tool. You can read more about all the fun stuff OpenAI did to make GPT 4.1 so darn powerful in this article, from the source.

Llama 4 wowed just about everyone

And let’s not forget Llama 4, the Open Source model from Meta that really blew some minds when it was announced the first week in April. Meta released Llama in a few different flavors, and it’s safe to say they’re a lot better at naming than OpenAI is. The core versions are called Scout, Maverick, and Behemoth.

If you want to do a deeper dive into Llama 4 and all is awesomeness, there’s a great article in Dev.to that I highly recommend.

xAI releases Grok Studio

There’s a new vibe coding platform in town, and it was literally just announced on Tuesday, so I haven’t had the chance to play around with it myself, but the vibe coding community on X is going bonkers about it.

This article from Tom’s Hardware is a great example of all the things you can do with Grok Studio, for free. As the title says, the writer made a website, research paper, and browser game, all in Grok Studio, and yes, all for free.

Chong has a great tweet above with an initial review and some videos of Grok Studio in action, and I’ll be honest, it looks pretty impressive. Since I haven’t had the chance to play around with it myself yet I can’t share how it compares to other tools, but you can bet that I’ll be diving in this week and seeing how it does.

Last but not least…is OpenAI buying Windsurf 👀

There’s a good chance you’ve already seen this as I seem to not be able to go anywhere and no read about it, but in case you didn’t - OpenAI is looking at potentially buying Windsurf for $3B. This is still a developing story but would be a pretty big win for both sides IMO.

Tutorial: how to build a full stack app in Cursor

I wanted to make sure to share at least one tutorial per issue for anyone who wants to learn how to vibe code by following a tutorial that walks them through it step-by-step. And this is a great one for those of you who might just be starting to dip your toes in the vibe coding waters.

Of course, a tutorial like this isn’t just for people who are new to vibe coding, there are some really good nuggets in this one, especially if you’re someone who understands frontend but not backend, or vice versa. I watched it myself and picked up a few nuggets so safe to say - if you watch it, you will too.

YC Partner Tom Blomfield vibe coded a hands free recipe site - RecipeNinja

Okay, this is just too cool not to mention. Like the title says, YC Partner Tom Blomfield vibe coded a site, but it’s not just a simple one shot and done project, it’s a pretty detailed complex platform, with a cool name - RecipeNinja.

Tom did a great job documenting his vibe coding journey on X, and I think this is also a great example of how vibe coding often requires using multiple models since Tom used quite a few, and generated 35,000 lines of code in the process, so yeah - not a small project.

For anyone who thinks Vibe Coding something useful means writing a prompt or two and poof, you’re done, I hope this shows that’s not really the case. There’s a lot that goes into vibe coding something actually functional and useful, and I think RecipeNinja is now one of the best case studies out there.

Congrats to Tom, and as someone who loves cooking myself, thanks for making this! 👨‍🍳

App Security rules every vibe coder should know

One of the areas that I find people often miss when it comes to vibe coding is security…and well, it might be one of the biggest mistakes you’re making right now. This has become a hot topic in the vibe coding world, and for good reason. If you’re vibe coding software, that you want other people to use, security needs to be on your radar because there’s no bigger way to lose trust, and face potential legal action, than a security breach.

Luckily CJ is lending a helping hand with this solid list of App Security rules. Of course, you might not know what everything on this list means, and that’s okay - for anything you don’t understand hop over to ChatGPT and ask, “what is ______” and just fill in the blank with whatever you don’t know.

Remember, vibe coding doesn’t mean not understanding what’s going on and hoping models do everything for you. Learn about app security, it’s worth the time.

Okay, and that’s a wrap thanks for reading the Vibe++ edition of the VibeCoding.sh newsletter. I hope you enjoyed this issue and as I said above, please, give me feedback, tell me things you like, didn’t like, what you want to see from this.

As a paying subscriber you should know that not only am I listening, I will take action with each and every suggestion I get. It is an incredible honor to have you onboard, thanks again from the bottom of my heart 🙏 🙏 🙏