In partnership with
Here’s how I use Attio to run my day.
Attio is the AI CRM with conversational AI built directly into your workspace. Every morning, Ask Attio handles my prep:
Surfaces insights from calls and conversations across my entire CRM
Update records and create tasks without manual entry
Answers questions about deals, accounts, and customer signals that used to take hours to find
All in seconds. No searching, no switching tabs, no manual updates.
Ready to scale faster?
Beginners in AI
Good morning and thank you for joining us again!
Welcome to this daily edition of Beginners in AI, where we explore the latest trends, tools, and news in the world of AI and the tech that surrounds it. Like all editions, this is human curated and edited, and published with the intention of making AI news and technology more accessible to everyone.
THE FRONT PAGE
AI Training Just Got 8 Million New Workers: DoorDash Pays Drivers to Film Everyday Chores
TLDR: DoorDash is paying its 8 million delivery drivers to film everyday tasks like washing dishes and folding laundry to train AI and robotics models.
The Story: DoorDash launched a new standalone app called Tasks that pays its delivery drivers to complete small assignments that feed data into AI systems. Examples include filming yourself washing dishes while wearing a body camera, recording conversations in another language, scanning store shelves, and even closing the doors on Waymo self-driving cars. The video and audio footage goes to DoorDash's in-house AI models and to partners in retail, insurance, hospitality, and tech. Since 2024, Dashers have already completed more than 2 million tasks. Pay is shown upfront and varies based on effort and complexity. The app is live in select U.S. markets but notably excludes California, New York City, Seattle, and Colorado, all places with stricter gig worker and data privacy laws.
Its Significance: Your delivery driver is now also an AI trainer, and this is part of a much bigger trend. Earlier this week, we covered how Niantic is turning Pokémon Go players into data collectors for its AI mapping models here AI + Pokemon Go? Pokémon Go Players Unknowingly Built a 30 Billion Image Database Now Training Delivery Robots . Now DoorDash is doing the same thing with its drivers. The pattern is clear, and unfortunate: companies are finding ways to turn everyday human activity into AI training data from as many sources as possible. Gig workers, gamers, shoppers, it doesn't matter. If you're moving through the physical world, someone wants to record it and feed it to a model. The excluded states tell you something too: where worker protections are stronger, this kind of program doesn't launch. Expect more of this, not less.
QUICK TAKES
The story: Amazon is working on a new smartphone codenamed "Transformer," led by Xbox co-founder J Allard. The phone would put Alexa at the center and make buying from Amazon and using Prime services easier than ever, and it might even skip traditional app stores entirely.
Your takeaway: Amazon's last phone, the Fire Phone, flopped hard over a decade ago. This time they're betting that AI assistants have gotten good enough to justify a phone built around one. Whether it works or not, it shows how tech companies see AI as the new front door to your wallet.
The story: OpenAI confirmed it's combining its ChatGPT app, Codex coding tool, and Atlas web browser into a single desktop application for Mac. The move is being led by Fidji Simo, OpenAI's head of applications, who told employees the company had been spreading itself too thin across too many products.
Your takeaway: Instead of juggling three separate OpenAI apps, you'll get one that does it all. OpenAI is trying to become the single hub for how you chat, code, and browse, which puts it in direct competition with Google and Microsoft for controlling your desktop.
The story: Cloudflare CEO Matthew Prince said at SXSW that AI bots will generate more internet traffic than humans by 2027. The math is simple: a person shopping for a camera visits maybe five sites, while an AI agent doing the same thing could hit 5,000.
Your takeaway: Before generative AI, bots were only about 20% of web traffic. Now that number is climbing fast, and it means websites, businesses, and the infrastructure behind the internet will need to be rebuilt around serving machines, not just people.
TOOLS ON OUR RADAR
📎 Paperclip Free and Open Source: A brilliant orchestration framework that allows you to easily hire and manage artificial intelligence employees to completely automate your business workflows. (Alternative to AutoGPT)
🏢 *Marblism Paid: A ready made AI agent platform that provides a full team of ready to deploy artificial intelligence employees to handle your inbox social media and lead generation with a special ten percent discount available for all new users.
Pumble Freemium: A highly organized team communication platform that offers unlimited message history and dedicated channels to keep your entire company perfectly aligned.
🎙️ *BetterDictation Freemium: A lightning fast voice to text application running directly on your Apple computer that flawlessly transcribes your spoken words into any software you use.
TRENDING
Feds Charge 3 in $2.5 Billion Scheme to Smuggle AI Servers to China - Federal prosecutors charged three men linked to Super Micro Computer with using fake documents, shell companies, and dummy servers to illegally ship $2.5 billion worth of AI servers to China. They even used hair dryers to peel off serial number labels and stick them on fake servers left behind for inspectors.
Adobe Firefly Now Lets You Train AI on Your Own Art Style - Adobe launched Custom Models in public beta, letting creators upload their own images to train a personal AI that generates new work in their exact style. It's optimized for characters, illustrations, and photography and costs 500 credits per training run.
White House Releases National AI Framework, Wants to Override State Laws - The Trump administration issued a six-part legislative framework pushing for a single federal AI standard that would block states from passing their own rules. It covers child safety, data center energy, intellectual property, and preventing AI censorship of political speech.
ServiceNow CEO Warns Gen Z Could Face 30% Unemployment From AI - ServiceNow CEO Bill McDermott told CNBC that Gen Z unemployment, currently at 9% for new graduates, could spike into the mid-30s within a couple of years as AI agents take over entry-level work. He predicts 3 billion digital agents will be added to businesses by 2030.
Google DeepMind Creates a 10-Trait Framework to Measure Progress Toward AGI - DeepMind published a "cognitive taxonomy" that breaks general intelligence into 10 measurable categories, from perception and memory to problem-solving and metacognition. They also launched a $200,000 Kaggle hackathon to build real tests for each one.
OpenAI Aims to Build a Fully Automated AI Researcher by 2028 - OpenAI's chief scientist told MIT Technology Review that building a fully autonomous AI researcher is now the company's "North Star." The plan: have an AI "research intern" by September 2026, then a multi-agent system that can independently tackle large scientific problems by 2028.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
✈ Build a travel packing list that organizes items by category, tracks what’s packed, and flags essential items
Build a travel packing list app. Features: add items with category (Clothes, Toiletries, Documents, Electronics, Medicines, Other), mark essential items with a star, check off packed items, packing status showing total/packed/remaining, essential items warning when unpacked, progress bar, category breakdown, filter by All/Unpacked/Packed. Dark teal theme. 900px wide.What this does:
Add items by category and mark the ones you can’t forget as essential. Check them off as you pack and the progress bar fills. The app warns you if any essential items are still unpacked when you’re almost done.
What this looks like:
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Visit 1,000x more websites than a human to complete a single task, generating enough traffic to overtake all human internet use by 2027.
❌ Still Can't: Tell the difference between a useful website and a junk one the way a human can, which is why those 5,000 bot visits per task create real server load with no guarantee of better results.
✅ AI Can Now: Turn a plain English description into a full working app with databases, login systems, and multiplayer features, no coding required.
❌ Still Can't: Independently design, test, and validate a scientific research project from start to finish. OpenAI says that's still at least two years away.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
AI researcher Janelle Shane collects the funniest, weirdest failures of artificial intelligence and uses them to explain how machine learning actually works. The title comes from an AI's attempt at a pickup line, and it gets better from there. It's the most fun you'll ever have learning why AI is simultaneously impressive and incredibly stupid.
“AI is Going to Fundamentally Change…Everything”
That’s what NVIDIA’s CEO said, calling AI “the largest infrastructure buildout in history.” Their chips helped make it happen. Now they’re collaborating with Miso Robotics for key robotics advances. Miso’s restaurant-kitchen-AI robots logged 200K+ hours for brands like White Castle. And NVIDIA helps unlock up to 35% faster performance. 100k+ US fast-food locations are in need, a $4B/year opportunity for Miso.
This is a paid advertisement for Miso Robotics’ Regulation A offering. Please read the offering circular at invest.misorobotics.com.
Thank you for reading. We’re all beginners in something. With that in mind, your questions and feedback are always welcome and I read every single email!
-James
By the way, this is the link if you liked the content and want to share with a friend.
Some * designated product links may be affiliate or referral links. As an Amazon Associate, I earn from qualifying purchases. This helps support the newsletter at no extra cost to you and Amazon makes a tiny hair less.
