Last week of March, first week of April. The shift is accelerating.
🗞️ Industry News: What Happened This Week
This week it's more about the AI companies — the parents — deciding where they should let their models go to school.
OpenAI's "Spud" Pivot
Rumors are swirling that OpenAI is winding down some of its video-heavy resources — including parts of Sora — to go all-in on a next-gen model codenamed "Spud." The focus? Agentic capabilities.
What that means for you: gig work is actively shifting from simple text-rating to complex, multi-step software interaction training. If you've been doing basic chat evaluations, the ceiling is about to drop on those rates. If you can guide an AI agent through a multi-tool workflow, the ceiling just went up.
Scale AI Launches "Scale Labs"
In late March, Scale AI officially launched Scale Labs — a new division moving beyond standard labeling to focus on evaluating advanced, multimodal AI in real-world environments.
Translation: a likely surge in projects requiring specialized domain expertise (legal, medical, advanced coding) rather than generalist tasks. If you have a verifiable professional background, this is the shift you've been waiting for. If you've been coasting on creative writing tasks, start building toward something more specific.
Anthropic's Economic Index
Anthropic released a new report showing that Claude usage is shifting toward "low-wage personal queries" and "digital collaboration." The read between the lines: the market is stabilizing around a new baseline where high-quality, nuanced human feedback isn't the exception — it's the requirement.
The floor is rising. So is the standard.
🔍 Platform Breakdown
I had to move away from a lot of my data-driven vision, so things are changing a bit. I did add a monthly insight page for each platform. The links are in the Quick Comparison table on the site. Those pages will be updated monthly with any insight I can gather into each of the platforms.
 DataAnnotation |  Outlier AI |  Alignerr |  Handshake AI |  Babel Audio |
 Stellar AI |  Mindrift |  OneForma |  RWS (TrainAI) |  Appen |
🛠️ New Tools for the Hustle
Instant Global Payouts (Thunes): On March 30, fintech giant Thunes launched a direct-to-workforce payout solution — instant cross-border payments into bank accounts or stablecoin wallets in 140 countries. If you've been grinding on international platforms and waiting 5–7 business days to see your money, this is the kind of infrastructure change that makes the whole thing more viable.
AI-Powered Tax Help (Eazy-C): A new tool called Eazy-C was announced for a 2026 launch. It's designed specifically for gig workers — automatically syncing with your bank accounts to categorize platform income and deductions (home office, data costs, equipment). If you've ever stared at a 1099-NEC from four different platforms and wanted to throw your laptop out a window, this is the tool you've been waiting for.
👀 Looking Ahead: The Skills Earthquake
We already talked about AI going to college, and you might be trying to figure out what these changes mean for you. At its most basic, this means you will not continue to get work if you don't have a technical understanding of how AI works. Prior to this, anyone could come in and find something to do — they didn't need to understand the task to complete it. Moving into chain-of-thought work requires you to actually understand what's being asked of you. Technical AI fluency is no longer optional.
Knowing how to prompt, debug, and guide an agent through a multi-step task is becoming the core requirement for the highest-paying gigs. Not a nice-to-have. Not a differentiator. The baseline.
The platforms that are thriving right now — Mindrift, Stellar AI, the upper tiers of DataAnnotation — all have one thing in common: they're paying a premium for people who understand how AI agents actually work, not just people who can rate whether Response A is better than Response B.
If you've been doing generalist work, the next 90 days are a good window to start building toward something more specific. Pick a domain. Get better at prompting for that domain. The floor is rising. The workers who move first are the ones who get to set the rate.
When people get together and discuss the projects they're working on, someone always says "How do I get on that project?!" — and 99% of the time, the answer is the same: I don't know. If it's at a company you're already contracted with, make sure your skills and resume show you can do that work. If it's a company you don't work with yet, get a link or a referral and apply for that specific project. Most of the time it really is luck — hopefully you're lucky. You never know when a unicorn will find you.
📈 Stay Ahead
Breaking Even tracks AI gig platform health by analyzing publicly available data sourced from community forums, message boards, and online discussions. Data is reported independently of any platform mentioned. Some application links may be referral links that pay a small commission for successful signups — this has no impact on what I report and there is no cost to you. Anyone who asks you to pay money for the opportunity to work or gain access to work should not be trusted.