CULTURE

DoorDash offers cash for AI data. Read the terms

A company is willing to pay people to capture real-world content to train robots. No, that’s not a Black Mirror episode, it's DoorDash’s new app. 

On Thursday, DoorDash unveiled Tasks, a feature that pays Dashers to provide businesses with ground-level insights, from photographing restaurant food to capturing hotel entrances for easier deliveries. Win-win, right? Well, tucked below that is a paragraph about a new standalone app DoorDash is piloting. 

“We’re also piloting a new standalone app where Dashers can complete activities like filming everyday tasks or recording themselves speaking in another language,” said DoorDash. “This data helps AI and robotic systems understand the physical world.” 

Real-world data is one of AI's most valuable and limited resources. Without it, manufacturers are stuck training models on simulations that only approximate reality. That's likely why DoorDash is willing to pay upfront: it's a premium commodity they can upsell. DoorDash has not responded to a request for comment. 

While getting paid to perform and record routine tasks may sound like a good way to make extra cash, users should soberly consider the privacy and security implications, which, according to Andrew Shimshock, CTO and Co-Founder of Mill Pond Research, a company focused on secure AI solutions, are numerous.

“Once an AI or robotic system is trained on biometric or spatial data and released into production, the underlying training data must be considered compromised,” said Shimshock. “Bad actors do not need to breach a server; they simply need access to query the AI.”

Shimshock also highlighted that once a person’s biological markers are baked into a multimodal AI, that data cannot be unlearned or deleted. Beyond the consenting individual’s data, there is also the question of what that means for individuals incidentally captured in the background of a task, without their knowledge or consent. 

While dystopian, this approach isn’t new. Data from the popular Pokémon Go game, which had hundreds of millions of players wandering the streets to capture the game characters, has now been used to create a Visual Positioning System (VPS) to train delivery robots. 

The lines of user consent are getting increasingly blurred in the era of AI. The question is, when is too much? It is an accepted truth that the entirety of the internet now lives in the datasets behind these powerful models, but what if our physical existence gets baked into them too?

DoorDash has long offered an accessible path to income, with few barriers to entry and real earning potential. That makes it genuinely valuable, which is why I worry that placing a comparatively easier task option alongside traditional deliveries could lead people to default to it without fully understanding the trade-offs. It is also worth considering that DoorDash likely stands to profit far more by batch-selling this data upstream, raising the question of whether contributors would be better served selling directly to buyers, if they can find the opportunities. But as long as supply meets DoorDash's demand, there is little incentive to cut out the middleman, and that is what makes this so difficult to untangle.

TOGETHER WITH MERGE

How teams plan to use MCP this year

Most teams building AI agents plan to adopt the Model Context Protocol (MCP) this year. Most of those same teams have serious security concerns about it.

To understand how teams are navigating this tradeoff, we surveyed hundreds of AI leaders building AI agents for the first-ever state of agentic integrations report.

Their top concerns?

• 70% worry about credential leaks and malicious servers

• 56% say MCP doesn’t support enterprise search well

• 51% report ambiguous tool definitions causing incorrect tool calls

Get your free copy to learn more.

PRODUCTS

Claude Cowork’s new Projects power up workers

Anthropic's Claude Cowork has maintained steady momentum since its January launch, with no signs of slowing.

On Friday, Anthropic launched a Project feature for the Claude Cowork desktop app and it allows users to create different workspaces that contain the tasks, instructions, and files relevant to that one area of focused work. 

The content in the Project is then stored locally in a folder on the user's computer. This has several advantages. Users never need to reupload the same information between sessions, can access it without constant context-switching, and retain full control over their files.

It also helps with vendor interoperability. Often, users don’t want to switch between AI tools simply because of the time it took to tailor them to their preferences and needs. With the instructions and files being stored locally, users can then easily reupload the personalization information elsewhere. 

To create a new project, users can either start from scratch by setting up a new folder with instructions and files, import from chat by bringing over instructions and files from an existing chat project, or use an existing folder they already work from. Anthropic said users can update or download the Claude desktop app to try it. 

While this is in itself a major update, it doesn’t seem like Anthropic is stopping there. The company is reportedly developing a Sketch tool for Claude, according to an X post. This would allow users to input a quick sketch to explain what they’d like done, rather than having to describe it in words.

The comment section for the project's release featured many memes. However, collectively, the memes were positive, highlighting the rapid pace at which Anthropic keeps shipping new features. Beyond being quite funny, it does highlight a bit of Anthropic’s strategy. The company seems to be laser-focused on its target audience and, as a result, constantly shipping features that improve their workflows. With both Projects in Cowork and the Sketch tool, Claude keeps getting progressively deeper into knowledge work, likely catering to its mostly enterprise clientele.

TOGETHER WITH ASAPP

The service operating model is changing

The agentic AI era isn’t coming. It’s here.

As AI agents resolve more customer intents autonomously, the contact center operating model changes. Human roles move from handling volume to governing, optimizing, and improving AI performance. Scale no longer depends on headcount.

Tomorrow, special guest speaker Kate Leggett, VP & Principal Analyst at Forrester, joins ASAPP on what this structural shift means for CX leadership — and how organizations operationalize mature human–AI collaboration.

If you’re rethinking how service is designed, staffed, optimized and scaled, this session is for you.

Join us live →

CONSUMER

3 AI voice tools our team recommends

As an AI reporter who uses transcription services daily and has used these tools for years, I'm well-qualified to answer the question I get asked most: which AI transcription tool do you use? Like most things, the answer is that it’s complicated. 

AI's strength in understanding natural language made transcription a natural fit, prompting many companies to launch solutions with unique twists, making it hard to choose the best one. So I'm going to break it down into a few categories. 

Best for transcription: Otter.ai  

Otter.ai does everything you’d want from a transcription tool, including offering accurate transcriptions, timestamps, summaries, audio playback, and even the option to ask questions about your transcript via a chatbot interface that uses LLMs. You can also import recordings, record live, or send your Otter bot to attend meetings and take notes. 

The one thing to be wary of is that you will want to pay for Otter.ai Pro ($100 annually), because the free version limits users to three lifetime imports and even paywalls access to older conversations that exceed your 25 limit, placing them in what I call Otter.ai jail. However, Otter.ai has been a fan favorite long before the AI craze, and it's well deserved. 

Best for meeting notes: Granola

Granola syncs with your calendar and joins meetings without disruptive bots, combining AI-generated notes with your own for polished results, and transcripts are impressively accurate. The chatbot portion is very helpful, too, as you can chat with it live during the meeting and after. The main drawback is the lack of audio playback, which makes fact-checking quotes tricky, though it's likely ideal for students. I’d recommend trying it for free, and scaling up if you like it and need more. 

Best for productivity: Wispr Flow 

If you have ever wished you could type as fast as you could speak, Wispr Flow allows you to do just that. While this tool serves a very different purpose from the rest of the list, it is just as handy. All you have to do is click a keyboard shortcut, speak into it, and watch your words appear accurately in any textbox. It's really that simple. Similar to Granola, I recommend giving the free tier a try. That’s what I use, and I haven’t run into any limits yet.

Do you have any other questions about AI products, whether hardware or software? Let me know via email, Instagram, Threads or X.

LINKS

• OpenAI slashes its compute spending target from $1.4T to $600B

• Super Micro cofounder indicted by U.S. for smuggling GPUs into China

• Amazon is reportedly building an AI smartphone codenamed 'transformer'

• Tesla and SpaceX announce Terafab to manufacture chips for robots and AI

• Tencent announces ClawBot as a claw agent integrated directly into WeChat

• Commentary: The race for power behind AI’s utopian story

• DuckDuckGo: The company just introduced a new Pro subscription plan catered to AI superusers with the highest reasoning on DuckAI, access to Claude Opus 4.6, and more

• Claude Code: Anthropic’s agentic coding platform now features channels that allow users to control the Claude Code session using Telegram and Discord via MCPs 

• Gamma: The AI-powered content engine released a series of updates, including Gamma Imagine, Connectors, and AI-native templates

• Goodnotes: A new integration between the AI-powered notetaking application and ChatGPT allows users to turn ChatGPT chats into Goodnotes documents

• Adobe: Principal AI Technologist

• World Economic Forum: Lead, Frontier AI Systems, Centre for AI Excellence

• Anthropic: AI Productivity Lead, Communications

• Palo Alto Networks: Distinguished Engineer - AI-First Transformation

GAMES

A QUICK POLL BEFORE YOU GO