🧙🏼 AI-powered app builders

Howdy, wizards.

Elon's $97.4B bid to buy OpenAI—and the tweets & drama that ensued—dominated headlines in AI this week. However, I want to put the focus on a couple of other fascinating topics that deserve attention.

[But, if you're still curious about “the tech industry’s Kendrick and Drake”, I've included some good links about the Musk-Altman beef at the end of this email.]

Here’s what’s brewing in AI.

DARIO’S PICKS

1. Anthropic drops an economic index and a lot of data on how AI is used

The Anthropic Economic Index aims to show AI’s effect on the labor markets and economy over time. It’s essentially a massive, open-sourced dataset of Claude conversations that researchers can use for studying and sharing findings.

Anthropic’s own initial report is based on millions of anonymised Claude conversations and shows how AI is used for real-world tasks by users across industries.

Some overall findings:

• AI adoption is highest for coding and writing/editing tasks

• 1/3 of jobs use AI in 25% of their tasks; Only 4% use AI for 75% or more of their tasks.

• AI is used slightly more for augmenting tasks (ie where you collaborate to get the final outcome) than for automating tasks (where you one-shot your way to a result). Augmentation accounted for 57% of tasks while automation was 43%.

• AI is used most in mid-to-high wage jobs (e.g. programming, data science). For the lowest and highest ends of the spectrum, usage is less. This likely reflects the limits of AI capabilities and the barriers for using new tech.

Top tasks by industry:

• Computer & math (37.2%): build/maintain apps, debug code, manage databases

• Arts & media (10.3%): produce creative content, handle PR, run marketing

• Education & library (9.3%): create curricula, teach subjects, organize publishing

• Office & admin (7.9%): routine IT, customer support, document data

• Science (6.4%): academic research, experiments/analyses

• Business & financial (5.9%): financial analysis, investment advice, ops metrics

‎ Why it matters‎ ‎ This is a great initiative and helpful to society. AI companies obviously want to know more about how people use their product — they’ll get a lot of free insights. Researchers, policymakers and businesses will be more able to make evidence-based decisions based on real-world data.

Keep in mind about this data is Claude-only. ChatGPT is probably more popular in other fields than Claude, so the data isn’t necessarily representative for AI usage overall.

TOGETHER WITH HUBSPOT

Need a personal assistant? We do too, that’s why we use AI.

Ready to embrace a new era of task delegation?

HubSpot’s highly anticipated AI Task Delegation Playbook is your key to supercharging your productivity and saving precious time.

Learn how to integrate AI into your own processes, allowing you to optimize your time and resources, while maximizing your output with ease.

Get the free guide here.

DARIO’S PICKS

2. Market map: AI-powered web app builders

a16z’s Justine Moore is back with another map of emerging markets in AI, this time of app and website generators.

These products use an LLM to generate code from your text prompt, process it through middleware that deals with files and API calls, then execute that code in a sandboxed environment and giving you a working application.

a16z’s research shows they’re good for simple builds and might feel like magic to non-technical users. However, they struggle as the apps get more complex—integrations and bugs are not easy to fix once a certain level of complexity is reached, and you might end up wishing you’d coded it from scratch.

The users of these tools span the entire range of coding and design skills:

• Entirely non-technical people are typically leveraging these tools to build small apps that cater to unique interests (like a bedtime story maker).

• Developers are using text-to-web app tools for faster prototyping and shipping new features.

• Consultants and agencies, which often get tasked with building simple landing pages for SMBs, are also using more AI powered options

Going forward, a16z expects app and website generators to become more niche, getting better collaboration features (for enterprise adoption) and having packaged integrations (authentication, payments, databases). Beyond app generation, a16z sees the biggest potential for companies able to go more vertical and help users through the entire workflow of creating, building and running an app.

If you’re curious to check out some of the most well-known app generators, here’s some links for you:

Website generation: Butternut AI, Dora, Dorik, Durable, Framer, Hostinger, Jimdo, Squarespace, Wix, 10web

Web app generation: Bubble, Databutton, Ohara, softr, val town,

Web app generation w/ code migration: bolt, co, create, heyboss, lazy, lovable, Replit, softgen, Tempo labs, Trickle, v0

‎ Why it matters‎ ‎ You might’ve already heard of AI tools like Cursor, Replit and v0 rapidly making web and app development accessible to less technical audiences (as well as boosting productivity for those who code). However, as most emerging markets, there’s many other companies fighting for a piece of the pie—and a16z does a solid job at digging up and showcasing those lesser-known, emerging players.

Keep in mind, a few of the examples above existed well before GenAI was a thing (e.g. Wix)—but they’re included as they have introduced features that make text-to-app/website possible.

QUICKFIRE HEADLINES

Side notes

Here are the other essential updates on AI you shouldn’t miss this week, that I didn’t have time to cover in-depth.

• Morning Brew said it best: Sam Altman and Elon Musk are the tech industry's Kendrick and Drake rn. TLDR: Elon Musk offered to buy OpenAI for $97.4 billion; Sam Altman’s response on X was “no thank you but we will buy twitter for $9.74 billion if you want.” The offer comes as OpenAI is transitioning to a for-profit and Musk has already sued them a bunch of times. Sam then went on a Bloomberg interview saying “probably his whole life is from a position of insecurity. I feel for the guy.”

• OpenAI has canceled its o3 release in favour of GPT-5. Sam Altman posted a roadmap update for GPT-4.5 and GPT-5. GPT-5 will be a system that integrates a lot of OpenAI’s tech, including o3. It will have a tiered scheme with different levels of intelligence for free, Plus and Pro users. GPT-4.5 is going to be OpenAI’s “last non chain-of-thought model”. The biggest takeaway is that they are trying to streamline their models and tools into a more unified user experience; the model picker as we know it might be coming to an end soon.

• Anthropic is dropping a new model in the coming weeks, according to the Information (yes, still a bit rumour mill). The company has been quiet since its release of Claude Sonnet 3.5 over half a year ago, while OpenAI, Google and others have launched progressively strong models. Apparently, it’ll be a “hybrid” model that can switch between deep reasoning and fast responses (kinda same direction as ChatGPT is headed) and some reports say it outperforms o3-mini-high.

FROM OUR PARTNERS

Your Voice AI Agent, Just a Few Clicks Away

Need a calling assistant that handles routine tasks like lead qualification and booking appointments 24/7? Synthflow has you covered with a library of pre-tested AI Agent templates tailored to industries like real estate, healthcare, and more. These ready-made solutions let you launch quickly—and if you’re feeling innovative, you can design and sell your own templates to earn extra income!

Try It Today

THAT’S ALL FOLKS!

Affiliate disclosure: To cover the cost of my email software and the time I spend writing this newsletter, I sometimes link to products and other newsletters. Please assume these are affiliate links. If you choose to subscribe to a newsletter or buy a product through any of my links then THANK YOU – it will make it possible for me to continue to do this.