TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Tell HN: Life Working as AI Data Trainer

2 点作者 charlieyu128 天前
I’ve been working as a Data Annotator since 9 months ago on several platforms. Mostly for Mathematics.<p>Why I started: This is the only job I can find since I moved to UK two years ago.<p>Pros: I’m making some decent money and saved a bit. At least I can work now.<p>And learning new maths again 10+ years after graduation is actually fun.<p>Cons: Instability. No work is guaranteed. One week you could work every hour you can work and the next week, nothing.<p>Bad actors. A lot of companies are just agencies getting projects from big tech and try to squeeze as much as possible. A lot of shady things happen. A project paying with decent pay ended and the same project came back with a quarter pay per hour. A few were practicing “pay-per-accepted-task” schemes, which is absolutely illegal in their country for not paying for hours worked. The funny thing is, they don’t do that in the US, but once it is outsourced, they think they can do anything.<p>Good companies that treats the worker fairly with respect are few and far, though they exist. As a general rule of thumb, if it is not possible for you to find your team leader for help, it is one big red flag. If your team leader is rude when you ask questions that are essential for your work, that’s another red flag.<p>How I feel about the future of the industry: As AI grows, so does the need of data trainers. We can’t build AI products for humans without human feedback. Human trainers are there to stay.<p>The landscape could be very different in several years though. There are lots of grifters in the AI industry now. As AI boom experiences slower growth, hopefully they will be gone.

暂无评论

暂无评论