DSC Weekly Digest 20 July 2021

COVID-19 has been in the news again lately, for several reasons. In many parts of the world, the delta variant of the virus has been hitting hard in those areas where vaccination rates are low. Not surprisingly, these are also the areas where there is a broad mistrust of science and where local leaders have sewn that mistrust for their own political gain. This is in turn breeding resentment and denialism in those regions that have the potential to turn into a vicious cycle, exacerbating geopolitical tensions and widening economic inequality.

It is still possible, with the vaccine, to still get COVID-19 Delta. though the likelihood is much smaller and the effects (and transmissibility) of the virus considerably tempered. However, this does not mean that the virus (regardless of variant) has become less dangerous to those who haven’t been inoculated, and even those who have had COVID-19 are not necessarily immune to the delta variant, though at least some of the vaccines seem to be better at provoking a full spectrum response.

This is raising the specter of a forever virus, one that may very well take years to fully recede, as potentially dangerous mutations continue to build up in broad pockets of “low-science” regions. The longer that this goes on, the more that the very nature of work and society is likely to change. Already, companies that had begun to require employees to come back to the office full time are reassessing those policies in the light of rising case numbers, though at least at this time the potential of going back into a full lockdown mode seems unlikely – hospitalization and death rates are not rising as quickly as has been the case before.

One of the more intriguing aspects of the pandemic has been that it has forced companies into thinking hard about what exactly they do. The demand for more sophisticated AI systems is rising, but there are also multiple indications that the field of AI itself needs to evolve dramatically first, taking into account not just faster AIOps platforms but increasingly needing to integrate with contextual stores and provide more intuitive interfaces that can adapt dynamically (and automatically) to given needs.

These won’t come from better algorithms. Rather, AI itself needs to adapt more readily as a component within a broader matrix of services that organizations use. This will become especially important as work becomes increasingly virtual and distributed, and is a key reason why data scientists and modelers need to start thinking beyond the analysis and toward the applications that rely upon that analysis.

In media res,

Kurt Cagle
Community Editor,
Data Science Central

To subscribe to the DSC Newsletter, go to Data Science Central and become a member today. It’s free!