Digest #541 | Build with 20k+ others - Become a Pro member⁴⁴

Hey folks, I just uploaded the recording of my Manus workshop²⁹⁹ - the AI agent from China that builds apps, takes browser actions and has its own sandbox.

Let’s get to what’s up with AI today;

TLDR; inside today’s newsletter

• new Gemini features feel ahead of the curve

• hackathon for vibe-coded games

• chats and new builds from our community

• 8 tools for building and marketing your apps

• which AI models help you get better

• other AI news and some good reads

Want to partner with us? Click here⁴⁶.

 🔎 Trends & news

• Google is shipping 🔥 lately and the current batch of updates shines up the Gemini app. It has three major upgrades¹⁸⁶:

  • Deep Research in Gemini is now available to free users too. It also has a new model underneath, which does more than just summarizing web articles. It tries to reason over them. The reasoning can still work but the vast number of resources it searches over makes it stand out.

  • This new model is an upgraded version of Gemini 2.0 Flash Thinking (still experimental). It has better performance and a longer context window—allowing it to answer complex questions and use multiple apps like Search, YouTube and Notes in a single go.

  • Gemini has a new mode called Personalization²⁴⁹. It asks you to connect your Google search history with Gemini and uses it when you’re asking Gemini questions. For example: when I asked it a question about Slack’s API, it looked at my search history, figured that I was using that with Webflow’s API and tailored its answer based on that.

• o1 and o3-mini in ChatGPT can now use “code interpreter”⁹⁴ i.e. run Python code to perform analysis before replying. OpenAI is slowly integrating all the ChatGPT features with o-series models.

• Levelsio is organizing a vibe coding game jam²⁶⁹ - 7 days to build a game with 80% AI-written code. The jury includes Karpathy and the creator of ThreeJS.

• Are developers nervous about AI agents’ impact on their work? Recent Salesforce research shows about 9 in 10 developers are actually excited about AI’s impact on their careers, and 96% expect it to change the developer experience for the better. Read why¹²⁷ developers are bullish about agentic AI.*
  *sponsored

💬 Inside the community this week

Becoming pro⁹¹ comes with access to our Slack community that builds and learns together. Here are some chats from over the weekend:

• Nicholas shared his second project for smart Gmail filters. (link¹⁷⁵)

• Héctor asked for mapping API recommendations. (link⁴⁸)

• Gabor asked for tools to create reels and posts with fixed layouts. (link⁷²)

• Daniel shared a webinar on AI coding with security in mind. (link²⁴)

• Keshav shared prompting insight as he builds our askBB bot. (link³⁴)

• Vivek asked about automating podcast guest research. (link⁴³)

• Mike asked about creating a consultant-like chatbot. (link⁷⁰)

Join the community to browse through replies or ask your own questions. Become a pro member today⁹¹!

p.s. just uploaded the recap for the last two workshops: level up your meme creation game³⁸ and make AI your business partner⁴⁸. The recaps are free and you can access recordings if you’re a pro member.

⚙️ Top new tools

• Memex¹⁸⁵ - Vibe code with an AI that can actually ship. Turn chat into working software on any stack, right on your desktop. Free credits with code 'BB'.*

• Mirage by Captions¹⁸¹ - Generate UGC with unique actors, backgrounds, and natural expressions.

• Figma to Bolt²³⁷ - Put bolt․new in front of a Figma URL & start building pixel-perfect apps.

• Docsforge^98† - Turn your React components into structured, user-friendly help documentation.

• OpenTools²²⁵ - Access the open ecosystem of MCP tools with one line of code.

• AI tutor by roadmap.sh²⁰⁰ - Learn developer skills with personalized courses.

• Bee AI by IBM¹⁸⁰ - Discover, run and compose AI agents from any framework.

• Stateshift Alignment^139†: Find out how focused or chilled out you are. (vibe coded by Jonny’s team¹⁴)

More tools here →⁹⁹
*sponsored

 🧠 On my mind…

Do AI models make us better problem solvers, or just faster ones?

While watching a session from Steph Smith’s Internet Pipes last night, I learned about the OFFSET function in a spreadsheet—a perfect solution to a common problem. It made me wonder, would AI have told me about this solution? In general, does AI help you get better at solving unseen problems?

So I ran a little experiment. I shared the problem with various AI models (without any prompting techniques) and asked them to help me fix it.

• Only o1 and o3-mini-high acted like real experts—immediately highlighting two common solutions (including introducing the offset function).

• Other thinking models (Claude, Gemini, o3-mini) did hint at the offset function but buried the lede with other much more complex answers. I have to act as the expert to find what’s best for me.

• Others weren’t any help at all. Just lists of overcomplicated answers.

This is a single instance, and prompting might change the results but I feel this does hint towards the fact the “reasoners” or “thinking models” are better teachers. And it matters because when you’re vibe coding, you encounter new problems every day, ending up in the “you don’t know what you don’t know” trap. so next time you feel like it, try a new chat with these models.

— By Keshav²¹

 📜 Interesting posts

• Multimodal media¹⁸⁹ - Code is now content.

• How ProPublica uses AI¹⁰⁷ responsibly in its investigations.

• Why I’m feeling the AGI²⁰⁷?

• Building AI agents to query your spreadsheets¹⁹⁰ and databases.

• xAI acquires Hotshot⁷¹, a generative AI video startup.

• Mistral⁶² and Cohere⁴⁷ both have new models in the market—if you’re interested in local models.

📌 To-dos

That’s it for today. Feel free to hit reply and share your thoughts. 👋

Enjoy this newsletter? Please forward to a friend.

Want to join a community of AI builders? Become a Ben’s Bites member²⁸ and join our Slack (bonus: access to all our workshops and courses).

Want to advertise in this newsletter? Click here⁴⁶.