Wednesday’s AI Report
• 1. 🧭 OpenAI's disappointing delay
• 2. 💖 How to get your product to market with Lovable
• 3. 🌍 How this start-up made $3B in revenue with AI
• 4. 👑 Make your job 1000x easier with Runner H
• 5. ⚙️ Trending AI tools
• 6. 💥 Google destroys journalism?
• 7. 🧠 Trump builds AI with ex-Tesla engineer
• 8. 📑 Recommended resources
Read Time: 5 minutes
✅ Refer your friends and unlock rewards. Scroll to the bottom to find out more!
OpenAI's disappointing delay
🚨 Our Report
Around six months ago, OpenAI’s CEO, Sam Altman, admitted that OpenAI has been on the “wrong side of history” when it comes to open-sourcing its models (the last time they released one was in 2019). To change this, he announced that OpenAI was working on an open-source model—to be launched “in early summer” (ie. June)—that would rival the likes of DeepSeek’s R1 and Meta’s Llama models (both are open source). However, recent setbacks mean that this highly anticipated model will now not arrive until late summer, at the earliest.
🔓 Key Points
According to Altman, the open-source model needs a “little more time” as his research team “did something unexpected,” which he believes “will be very, very worth the wait.”
While nothing has been confirmed, experts are speculating that this new, open-source model will be able to connect with OpenAI’s cloud-hosted AI models, meaning it will have the ability to complete highly complex queries.
The industry has recently exploded with new, open-source reasoning models. Alongside DeepSeek and Meta, Mistral just released open reasoning models–Magistral—and Qwen launched a hybrid version in April.
🔐 Relevance
Despite the disappointing delay of its new, open-source reasoning model, OpenAI has just launched a “more capable” version of its (closed) reasoning model, o3, called o3-pro. o3-pro (which is available now for ChatGPT Pro and Team users) can search the web, analyze files, use Python, and personalize its responses by leveraging the memory feature. Even though o3-pro takes longer to respond, temporary chats have been temporarily disabled (technical fault), and it can’t generate images, reviewers reportedly preferred o3-pro over o3, especially in “key domains like science, education, programming, business, and writing help” and rated it “higher for clarity, comprehensiveness, instruction-following, and accuracy.”
Build, launch, and ship your product—this year
Got a product you want to build, launch, and ship to market?
Lovable Shipped is a 6-week program—starting Monday, June 16th—that gives you the tools you need to:
Build your new product
Launch it to investors and the media during a live demo day in San Francisco
Ship it to market—this year.
Sign up now and get exclusive access to:
Weekly lectures, insights, and mentorship from industry experts to help you learn, build, and ship faster
Networking opportunities, connections, and support from a community of like-minded product builders and entrepreneurs
$3M+ in partner credits and perks to get your product to market.
Applications accepted from Friday, June 13th.
How this start-up made $3B in revenue with AI
Fintech start-up, Affiniti, saw that many SMBs couldn’t access CFO-level insights, so were making reactive decisions and mismanaging finances.
They launched an AI-driven "AI CFO" platform, which used advanced LLMs and data analytics to manage finances and corporate card services.
The platform automatically tracked expenses, reconciled transactions, and produced key insights for SMBs.
As a result, Affiniti secured $17M in Series A funding to expand the platform and made over $3B in revenue across their SMB base.
Huge News in the world of AI Agents
Runner H is the last AI tool you'll ever need!
Backed by a historic $220M seed round, H just released their flagship AI agent that lets you automate entire tasks in a few seconds.
Some tasks you can delegate to Runner H:
Reading your important emails and drafting (or even sending) replies
Creating a Google Sheet with trending ad ideas and sharing it with your team on Slack
Finding job opportunities and applying on your behalf
Offload some work to Runner H and see how good it feels!
Just a year since Google launched its AI Overviews (which uses AI to summarize search results) and recently its “AI Mode” chatbot, a study has found that these AI features have severely reduced visits to news sites.
According to The Atlantic, “Google is shifting to an answer engine,” and news outlets must rethink their business strategy as web visits from Google are likely to drop to zero over time.
Click-free answers are a real threat to journalism: Business Insider has cut 21% of staff (due to a drop in site visits), the NYT has seen a 66.5% decline in organic visitors, and HuffPost has lost over half its organic visits.
According to leaked code found in developer platform, GitHub, the Trump administration is developing a platform and AI tools to “accelerate government innovation with AI” which is expected to launch on July 4th.
Former Tesla engineer, Thomas Shedd, is leading the project, and has previously discussed using AI to detect fraud, analyze government contracts, and create “AI federal coding agents.”
The new platform will host three tools,“powered by the leaders in American AI”: An AI chat assistant, an API to connect to OpenAI, Google, and Anthropic models, and a console “to analyze agency implementation.”
The AI-minded path to scalable and responsible innovation
This podcast dives into what successful enterprise AI adoption looks like and how to prepare a workforce for AI when automation changes the expectations of work quality and speed.
We read your emails, comments, and poll replies daily.
Until next time, Martin, Liam, and Amanda.
P.S. Unsubscribe if you don’t want us in your inbox anymore.