Google essentially pulled into first place in many of these AI categories from what's available right now. I think the most significant is the image and video generation, which is absolutely mind-blowing. The new Flow product that I reviewed generates video and audio at the same time, making it feel so different from any video generator that does video only. In many cases, you wouldn't be able to tell that the output is AI-generated.

I also want to spotlight a few more apps that I didn't have time to test in the original video. One is Project Mariner, a computer agent you can use through the browser or as a browser extension. In our testing, it works great on certain pre-trained tasks, but it can struggle outside of those boundaries. Another is Stitch, an app that builds interfaces from scratch for you. You can simply prompt it for the type of app you want, and it generates the starting point, including custom icons and layouts.

Additionally, I want to mention Prompt DJ, which uses the new LIA model. This tool allows you to add different instruments to a musical composition through prompts, creating a fun and interactive way to generate music.

Overall, Google has made significant strides in AI, particularly in the areas of image, video, and audio generation. These releases have the potential to revolutionize various industries and workflows.

The recent Anthropic event unveiled two groundbreaking AI models - Claude Opus 4 and Sonnet 4. According to the latest benchmarks, these models have instantly become the best coding models in the world.

The key highlights are:

• Unparalleled Coding Ability: Claude Opus 4 and Sonnet 4 have demonstrated exceptional performance on real-world coding tasks, outperforming all other language models.

• Practical Usability: These models excel at "one-shotting" applications, meaning they can generate complete, functional code from a single prompt. This makes them highly practical for developers.

• Exceptional Writing Quality: The models produce writing that is indistinguishable from human-generated text, with an "absolutely insane" level of quality.

Additionally, the models have the unique ability to critique and provide feedback on written work, making them valuable tools for editing and improving code. The models' strong adherence to prompts, even with long context, further enhances their usefulness.

These advancements in coding-focused AI models are poised to revolutionize the way developers work, offering unprecedented productivity and efficiency. As more applications integrate these powerful tools, we can expect to see a significant impact on the coding landscape.

I think there's one section in the HubSpot guide that's particularly useful - the big visual comparison table. This gives you a quick overview of the similarities and differences between the major AI models like ChatGPT, Claude, and Gemini. Having this kind of side-by-side comparison is the best way to quickly understand the strengths and weaknesses of each tool.

The guide also includes deep dive playbooks for each model, explaining where they shine and walking through sample use cases. This can help you determine which AI assistant is best suited for your specific needs, whether that's research, writing, coding, or something else.

Additionally, the guide covers how to "multi-thread" AI, using different models for different parts of a workflow. For example, starting research in ChatGPT, then using Claude for the actual writing, and Gemini to polish things up. This kind of AI orchestration can be really powerful.

Overall, if you're feeling overwhelmed by all the AI assistant options out there, I'd highly recommend checking out this free guide from HubSpot. It can help remove the guesswork and ensure you're using the right tools for the job.

OpenAI's Codex is a groundbreaking release that is set to transform the way developers work. This new standalone website, accessible through chatgpt.com, is exclusively focused on integrating with GitHub repositories.

Once you connect your GitHub account, you can select a specific repository and branch to work with. Codex then allows you to run multiple tasks in parallel, showcasing its unparalleled coding abilities. The quality of the code it generates is truly remarkable, often surpassing the work of experienced full-stack engineers.

One of the key features of Codex is its ability to work with existing codebases. Rather than creating new projects from scratch, Codex shines in tasks like adding new features, refactoring code, fixing bugs, or even rewriting entire applications. The virtual environment it provides allows you to manage dependencies and set up the necessary tools and libraries.

For example, if you want to add voice input functionality to your chatbot, Codex can generate the required code and documentation, which you can then easily merge into your main branch through a pull request. This level of integration with GitHub makes Codex a powerful tool for developers who want to streamline their workflows and leverage the power of AI-assisted coding.

While Codex is currently only available through the $200 Pro plan, OpenAI has plans to make it accessible through the $20 Plus plan in the near future. This democratization of AI-powered coding tools is a significant step forward, empowering both experienced and novice developers to enhance their productivity and creativity.

Overall, Codex represents a major advancement in the field of AI-assisted development, and its integration with GitHub makes it a game-changer for developers who want to harness the power of AI to improve their workflows and deliver high-quality code more efficiently.

One of the most significant announcements from Microsoft Build this week was the open-sourcing of GitHub Copilot. This AI-powered coding assistant, which previously required a subscription, is now available under an MIT license, allowing developers to freely access and build upon the technology.

The open-sourcing of GitHub Copilot is a game-changer, as it enables developers to integrate and customize the AI assistant to fit their specific needs. This move aligns with the growing trend towards more open and collaborative AI development, where developers can leverage and build upon existing tools and models.

In addition to the open-sourcing of Copilot, Microsoft also announced updates to their Copilot Studio, which provides a suite of tools and features to enhance the AI-assisted coding experience. These updates include improvements to the multi-agent workflows, allowing developers to seamlessly integrate different AI models and assistants into their development processes.

While the open-source nature of Copilot may not align with the interests of all viewers, it represents an important shift in the AI coding landscape. By making this technology more accessible and customizable, Microsoft is empowering developers to explore new ways of integrating AI into their workflows and potentially creating their own AI-powered coding assistants.

For those who are already using or interested in Copilot, these updates and the open-source availability provide new opportunities to experiment and potentially contribute to the ongoing development of this influential AI tool.

We ran into a new app called Lav Art that launched recently, and the most interesting aspect is its interface. It takes a similar approach to Canva, where you can create marketing materials with the help of an AI-powered bot.

When you interact with the bot, it generates content for you, such as posters, mockups, and entire marketing campaigns. The best part is that the generated content is fully editable within the interface, so you don't have to reprompt the bot to make changes.

Lav Art also features a massive canvas where you can create and manage your entire marketing efforts. This could be particularly valuable for those in the marketing space who need to quickly produce high-quality visuals and content.

Compared to the bigger players like Canva, Lav Art presents a fresh challenge, offering an AI-driven approach to creating marketing materials. If you're looking for an efficient way to generate and customize your marketing creatives, Lav Art might be worth your attention.

• OpenAI's Move into Wearables: OpenAI has acquired IO, a hardware company led by Johnny Ive. While details are scarce, this acquisition suggests OpenAI's interest in developing AI-powered wearable devices. The vision for an "AI subscription of your life" discussed by Sam Altman provides some insight into OpenAI's long-term plans in this space.

• Hicks Field AI's Ad Generation Feature: The AI video platform Hicks Field AI has introduced a new feature focused on generating advertisements. This could be of interest to those looking to leverage AI for their advertising needs.

• Future House's End-to-End Scientific Discovery: The research agents developed by Future House have demonstrated the ability to autonomously suggest experiments, analyze data, and generate new hypotheses - a significant milestone in AI-driven scientific discovery. This showcases the potential for AI to accelerate scientific progress.

• Fortnite's AI-Powered Darth Vader: The popular game Fortnite has implemented an AI-generated voice for the Star Wars character Darth Vader. While the initial implementation faced some challenges with profanity, this example highlights the growing integration of AI-powered voices in video games, a trend that is likely to continue.

This week has been a whirlwind of activity in the world of generative AI. From the avalanche of AI products and features unveiled at Google I/O, to the impressive benchmarks set by Anthropic's Claude Opus 4 and Sonnet 4, the pace of innovation is truly staggering.

The standout highlights include Google's Flow, which can generate videos with dialogue, music, and background sounds, and Anthropic's Claude models, which have set new standards for coding ability and writing quality. The ability of these models to adhere to detailed prompts and provide constructive feedback on writing is particularly noteworthy.

Beyond the major events, we also covered a range of other interesting developments, such as Microsoft's open-sourcing of GitHub Copilot, OpenAI's foray into wearables, and the integration of AI-powered voices in Fortnite. The rapid pace of progress in these areas is a testament to the transformative potential of generative AI.

As the landscape continues to evolve, it's crucial to stay informed and to carefully evaluate the various AI tools and platforms available. The guide from HubSpot, which compares the capabilities of ChatGPT, Claude, and Gemini, can be a valuable resource in navigating this rapidly changing space.

In the end, the key takeaway is that the world of generative AI is moving at a breakneck pace, and it's essential to stay on top of the latest developments to ensure you're leveraging the right tools for your needs. By doing so, you can unlock the full potential of these transformative technologies and stay ahead of the curve.