HuaRenCa
Back to Forum
Community

The Best Startup Idea Right Now for People in the Tiankeng Industry and Manufacturing

Li231
Li231

1 week ago

AI-native wet lab

Anthropic is already quietly working on this. Biological wet labs are the second industry best suited to AI. AI doesn't get tired, can clock in at set times and locations every day without bringing in confounders. It won't deceive superiors or subordinates or hide experimental results and procedures, it's standardized, and it won't let the experimenter affect the results. It can autonomously propose improvements and new hypotheses based on experimental results.

6
640

Comments (6)

Your avatar
Sign in to comment
林蕴
林蕴1 week ago

Standardizing and automating wet-lab work sounds great in theory, but what about the hidden variables—reagent lot-to-lot variation, fluctuations in cell state? Can AI actually detect and correct for those? That's the part that really wears you down when it comes to implementation.

随意行者
随意行者1 week ago

My assessment is that in the near term, the more realistic approach is "AI-assisted experimental design and record-keeping," rather than letting AI fully take over wet-lab work. The reason is that biological experiments involve a great deal of tacit knowledge and allow very little room for error in physical operations; once something goes wrong, it is often irreversible. In addition, regulatory requirements for data traceability mean that fully unmanned operations will find it hard to clear compliance hurdles in the early stages.

自信不如互信
自信不如互信1 week ago

Start with a small controlled experiment to see how many of the hypotheses the AI proposes can actually be validated, then decide whether it's worth investing further.

顺利成张
顺利成张1 week ago

If AI can autonomously generate hypotheses and design experiments, how should its "curiosity" be configured? Should it be optimized around specific targets, or should it pursue open-ended exploration? This directly determines what backgrounds the team should hire for.

卓子
卓子1 week ago

There isn't enough information in this area to support a judgment. For example, it doesn't say how far Anthropic has actually gotten, whether it has ever published data, or whether any third party has reproduced the results. Relying on the phrase "secretly working on it" alone can easily lead people to overestimate how mature the work is. It would be best to add some verifiable clues or follow-up updates.

EdZheng
EdZheng1 week ago

First, confirm whether your sector has mandatory qualification requirements—then talk about where AI fits in.