Hacker News new | comments | show | ask | jobs | submit login

I echo that sentiment. To share another testimonial, I work in a vaccine manufacturing facility with control systems and robotics. The only reason biologics manufacturing facilities haven't been built overseas with the same zeal as other factories is the stringent requirements by the FDA, MHRA, etc. on manufacturing facilities that use live cultures. Many of the technician jobs are similarly dedicated to pushing carts and transferring components to the robots for processing. Because of the reduced efficiency of old drug discovery methods, many of these jobs are taken by college educated science majors who are finding that jobs in drug discovery research are quickly dwindling.

However, I disagree with the idea that robots are going to be able to take over in a five year timeframe.

I agree with you that the capabilities for robotics to do these tasks exist. "The future is already here - it is just unevenly distributed." I doubt however whether large manufacturing companies would be willing to shoulder the huge investments needed to develop robotics for these applications even in a healthy financial climate. Maybe this type of development could come from a company that proposes to do consulting work for manufacturing facilities, for example, Honeywell.

Additionally, I've seen that while companies see automation as an asset in pharmaceutical manufacturing, they also view it as a huge liability, as a single automation error can lead to enormous problems that take time to detect and can have huge costs, while human error generally leads to small mistakes that can be corrected and detected easily. Not sure how it is in other manufacturing subindustries.

"However, I disagree with the idea that robots are going to be able to take over in a five year timeframe."

That's not quite what I said. I specifically meant that robots will be able to take over the task of moving things around in the five-year timeframe. That is, the guys moving bins of things from here to there. We've already seen warehouses adopting this technology with a fairly high degree of sophistication today, so it's not really a far-out prediction. I'm not predicting that something that doesn't exist will be created, I'm predicting that something that does exist will be commercialized to the point it will become irresistible.

Remember, even if the robots are initially expensive, to replace an employee for a year opens up about $50,000 in capital to play with, assuming a low-paid employee + insurance (yours, not the employee's) + overhead, give or take $20,000. $50,000 is becoming a lot of robot.

Completely taking over in five years, no, definitely not. Certainly not in the facility I saw. But the long-term job trend will be clear in five years.


I guess we both agree on the long term trend and at this point, that's all that really matters.

But the task of creating highly reliable industrial robots is really quite difficult and quite a bit more expensive than $50,000. We have 12 robots in our facility and the field testing and applications testing process for a similar facility elsewhere is taking approximately 2 years and counting. This doesn't include the entire development and specification process of developing the robots as well as the control systems. Each robotic system is upwards of $5,000,000.

A lot of this cost comes from the additional points of failure. Large systems are prone to failures due to small, cascading errors. This necessitates a larger testing and development period.

This is different I'm sure for other industries however. The cost of a failure is exponentially greater in pharmaceutical manufacturing as compared to other forms of manufacturing. Its possible that costs could be much lower in other industries where the cost of a failure is reduced.


Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact