M
@mathematicanese
How long before people just be start uploading AI videos for easy money.
3
0
661
@bearlyaiposted on X
DoorDash released a new app called Tasks. It allows the 8 million Dasher couriers in America to create video and audio content for AI robotics training data. The App Store page shows “load dishwasher”, “make bed”, “fold clothes” and “wash dishes” as chores. Payment is $5 and up (with video instructions) with another page showing an option to make a 20-minute “natural language Spanish” audio for $20.
View original tweet on X →
Community Sentiment Analysis
Real-time analysis of public opinion and engagement
Sentiment Distribution
85% Engaged
77% Positive
Positive
77%
Negative
8%
Neutral
15%
Key Takeaways
What the community is saying — both sides
Supporting
1
Monetising the linguistic demographics
DoorDash is packaging its largely Latino, native-Spanish courier base as a data product — offering paid tasks that extract valuable natural-language Spanish audio from workers and turning workforce traits into revenue.
2
Workers are effectively training their own replacements
Several replies frame this as labour feeding future automation, where gig workers generate the training material that will later be used to automate the same jobs.
3
Data is worth far more than the pay
Commenters call out the mismatch between the strategic value of Spanish conversational audio or household-task videos and the small per-item payments (e.g., $20/20min, $5/video), raising AI-ethics concerns about undercompensation.
4
DoorDash as a distributed AI data-collection infrastructure
Many see this as an infrastructure play — a stealthy, scalable way to build proprietary training datasets using an existing courier network rather than a simple product feature.
5
People will game it and some will eagerly participate
Others predict or admit that users will upload content for easy cash — a mix of opportunism (“I’m totally doing this”) and concern about low-quality or staged submissions changing dataset integrity.
Opposing
1
“That can’t be real” — outright disbelief
Replies insisting the post is staged or generated by a bot, asking for proof or calling it a fake/deepfake.
2
Impressed admiration
Fans marvel at how convincing it is, praising @bearlyai and treating the example as a breakthrough in realism.
3
Jokey escalation — “BFF?” / “wut next?”
Sarcastic takes imagining AI as a best friend, partner, or replacing everyday human roles — played for laughs or mild unease.
4
Ethics and privacy concerns
Critics warn about deception, consent, data use and the broader implications of AI passing as human.
5
Practical testers and skeptics
People asking for verification steps, prompts, demos or tutorials to reproduce the result or prove authenticity.
Top Reactions
Most popular replies, ranked by engagement
S
@shifty_stan
Literally training their own replacements.
3
0
631
T
@themasmedia
nobody: doordash product manager: "hope a robot can start trimming my plants someday!"
2
0
1.2K
B
@bowtiedwhitebat
cant be real gesus wut next? BFF ?
0
0
339
Report an Issue
Found something wrong with this article? Let us know and we'll look into it.