Your article is very enlightening about the benefits of predicting maintenance problems in rigs and it’s impressive how RDS has implemented these techniques in their major operations. As I think about this, I can’t help but wonder how an ML algorithm would fare in predicting large scale disasters, especially because such disasters have been few in number so data can be hard to find and might be triggered sometimes by the most random and seemingly insignificant factors. To me, it looks like there’s going to be the need to do some sort of balancing between the signals generated by machines and the intuition of engineers – how this can be done I cannot wait to see!
As I read this I wonder if innovation in this industry will shift from OEMs and large companies to “startups”. Given the capital intensive methods of innovation now, a lot of the new developments seem to be coming in from existing players. 3D printing, to me has the potential to allow smaller firms, or even individual designers to manufacture. Since the customers would be the OEMs, they would potentially have to do the testing and sign-off but it appears to be a good place to start!
I think one of the major challenges for this model is that manufacturing at scale of pharmaceutical products is tightly regulated (as it should be). The costs are also prohibitive for a newcomer, let alone a non-profit that is not attempting to patent its inventions. I agree with the comment above on the need for a partnership with the government that can potentially provide cost assistance when it comes to manufacturing, either by focusing only on the required regulations rather than an all-encompassing set or through a grants program. For molecules that are mandated by the WHO etc. (which might require a minimum fixed volume), perhaps CO-ADD could work with contract manufacturing organizations and negotiate a preset price at fixed margins?
This beautifully written essay highlights how in the future consumers may not have to pay through the roof for niche items! While mass production has allowed a lot of people access to essentials (and then some), the same techniques do not work for any goods that have a limited market. I agree that 3D printing has allowed companies like Gantri to function but I am curious what their competitive advantage will be especially since a competitor can also manufacture easily, especially if they do not employ the designers or own their designs. In the long term, if they evolve to become a platform that connects designers and manufacturers, I think they should continue to manufacture which would be their core competency but also operate a marketplace model for manufacturers.
Akash, this is a brilliant article about using research to drive further research! I think as investigators continue to rely on systems like this to tweak decisions in their iterative testing process, they can use their time to do more of the divergent thinking you talked about. My understanding of ML is that we are relying on machines to identify pathways to an acceptable convergence but in a field like this where sometimes divergent thinking provides solutions, I think there is always a role for humans. In addition, while Kebotix’s products would help reduce the impact of bias by choosing the best path forward rather than on intuition but at the same time, the researchers’ bias might be what would lead to a breakthrough!