Was discussing it with a friend who is very into the world of startups (has one himself). It is a step in the right direction, but I think that people should be careful about using it. The thing is, this will replace many repetitive jobs. Currently, those are the jobs of many Phd students. So PIs don't have to hire as many PhDs to do the job that needs to be done. Which is good and bad at the same time. If they see in it as a way to simply cut costs and not invest that into letting PhDs actually do what they should be doing, the thinking, then its bad. Additionally, some countries have lab technician as a job (not the U.S., but Germany and the UK for example). Those jobs could be endangered. But it could also be an opportunity for some technicians to do something more "exciting". I know at least 5 technicians whose work is doing the same qPCRs and ELISAs over and over again. I wonder what they think about it.
It's good to keep in mind that the service has different value in the context of industry vs. academic research. With this and a few of the other funded startups, I have a feeling YC is going to try to work its way into biotech by keeping as much of its technology in the hands of programmers / the virtual world and use these sorts of services as a base API layer on which they can script their own experiments. It's not that robots don't exist in the rest of the pharmaceutical industry, but this lets them begin other companies without a meter of lab space. Keep development costs and complexity low and you can fund many more crazy ideas. Now a failed drug doesn't lead to abandoned lab space. Sure, you can screen many more drugs in parallel on this platform, but those exist already and it's probably that high-throughput isn't the rate-limiting step for progression on all fronts of biotech. ---- The other front is as what was already contracted services to academic labs. This has existed for decades in both the form of robot time (DNA synthesis / sequencing) and technician time. The upsides here are (hopefully) improved repeatability of experiments and a one-stop shop for all of your needs. And maybe more parallelized work. And maybe less time wasted feeling like you could've been replaced by a trained monkey. But there too, cost is often a limiting factor. Along with reagents, labs would then be paying for machine and storage time, which is cheaper to do in-house when you're already paying for the labor and you've already bought the freezers. So I think there will be a lot of inertia before academic labs begin picking up these sorts of services. Plus, a trained postdoc is almost always going to be more flexible than a lab robot. They'll have the expertise and eyes to make sure they don't pipette the pellet of cell lysate. They'll be able to find and set the optimal temperature for cell growth, rather than picking incubation options from a pre-defined list. They'll be able to ensure a folding experiment follows the exact time points defined in the protocol. It's not like a robot couldn't do all of these things eventually, but the early offerings of these services aren't really inclusive of the range of experiments a human-operated lab can do. ---- Long term though, I hope that researcher jobs transition more into the realms of method development and data analysis. There are vast sets of -omics datasets lying about that are often amenable to further analysis (All of my undergrad research was spent on previously published PDB structures). And most protocols across all fields of molecular and cellular biology are far from optimal. Right now, most researcher time is spent setting up or carrying out these experiments. In an ideal world, once a new super-resolution microscopy experiment was published, others would be immediately be able to use it for their assays. Once a new method for genome assembly was published, you could immediately begin using for new synthetic organisms without changing your code.