The idea of indestructible killer robots may sound like something straight out of the Terminator movie. But they could soon become a reality, as scientists have just witnessed metal healing itself for the first time, without any human intervention. A US-based study has overturned everything we thought we knew about metals by revealing that cracks from wear and tear can actually mend themselves under certain conditions. It's a discovery that has the potential to revolutionise engineering, with the prospect of self-healing engines, planes and even robots now on the horizon. 'This was absolutely stunning to watch first-hand,' said Brad Boyce, a scientist at Sandia National Laboratories who led the study with Texas A&M University.

Welcome to The TechCrunch Exchange, a weekly startups-and-markets newsletter. It's inspired by the daily TechCrunch column where it gets its name. Should early-stage founders ignore the never-ending debate on server infrastructure? Up to a point, yes: Investors we talked to are giving entrepreneurs their blessing not to give too much thought to cloud spend in their early days. But the rise of machine learning makes us suspect that answers might soon change.

Bare metal servers can also be specialized for a type of use case, for example by adding a GPU to accelerate computation for AI and machine learning …

Nvidia and VMware today announced the launch of an accelerated GPU service on VMware Cloud on AWS, forming an advanced hybrid cloud infrastructure for machine learning workloads. The new service will allow organisations to migrate VMware vSphere-based applications and containers to VMware Cloud on AWS, VMware's cloud service that runs on bare metal infrastructure in AWS data centres, where they can take advantage of high-performance computing, machine learning, data analytics and video processing applications, backed up by Nvidia accelerators. Specifically, VMware Cloud on AWS customers will be able to rent Amazon EC2 bare metal instances, an AWS service that provides resizable compute capacity, that are accelerated by Nvidia T4 100 GPUs. The lynchpin of the hybrid platform is Nvidia's vComputeServer, virtual vGPU technology that enables GPU-accelerated deployment of workloads in virtual environments. The technology is not technically new but Nvidia has expanded support to VMware virtual environments including vSphere, vCenter and VMware cloud.

Performance of Machine Learning workloads using GPUs is by no means compromised when running on vSphere. In fact, you can often achieve better aggregate performance, i.e. throughput of many jobs, by running on vSphere vs. bare metal A key benefit of running GPU-based Machine Learning workloads on vSphere is the ability to allocate GPU resources in a very flexible and dynamic way. Performance of Machine Learning workloads using GPUs is by no means compromised when running on vSphere. In fact, you can often achieve better aggregate performance, i.e. throughput of many jobs, by running on vSphere vs. bare metal A key benefit of running GPU-based Machine Learning workloads on vSphere is the ability to allocate GPU resources in a very flexible and dynamic way.

These days, organizations are creating and storing massive amounts of data, and in theory this data can be used to drive business decisions through application development, particularly with new techniques such as machine learning. Data is arguably the most important asset, and it is also probably the most difficult thing to manage. It can be structured or unstructured, and it is increasingly scattered in different locations – in on-premises infrastructure, in a public cloud, on a mobile device. It is a challenge to move, thanks to the costs in everything from bandwidth to latency to infrastructure. It has a zillion different formats, sometimes chunks of data are missing, and usually it is unorganized and alarmingly often ungoverned.

I begin with an assumption that you are somewhat familiar with "whitebox" term as it relates to networking gears and cognizant of its trend. The term "whitebox" and "open networking" to some extent synonymous, both innately suggest networking gears that are "open" meaning adheres to "open" standards and ecosystems. The former is a byproduct of open networking concept and refers to the disaggregation model in which hardware and software are separated. This decoupling of hardware and software created opportunities for a software ecosystem to flourish: well, almost! As with any technologies advents, early days are little bumpy.