Tech News Review week 9-17 November 2020, by Stelios Iordanis


DOD Pours Millions of Dollars into Print-on-Demand Drugs – Defense One

These systems make drugs from starting materials to finished product in hours instead of weeks or months. Critical active pharmaceutical ingredients (API) and their starting materials have long been sourced from overseas, such as China and India and the USA consider finding drastic solutions for breaking this dangerous dependence.


Synthetic Nanobodies, “Sybodies,” Found That Neutralize SARS-CoV-2

Nanobodies, small antibodies found in camels and llamas, are promising as tools against viruses due to their high stability and small size. Although obtaining them from animals is time consuming, technological advances enable rapid selection of synthetic nanobodies, called sybodies. A technology platform to select sybodies from large synthetic libraries was  developed that can provide one more weapon against COVID-19 and future viruses.


Separate Antiaging Trials for Rapamycin and Metformin |

AgelessRx claims that PEARL is the first nationwide telemedicine trial and one of the first large-scale intervention trials on Longevity. They will evaluate the safety and effectiveness of rapamycin in 200 healthy adults for Longevity; rapamycin is still a very expensive antibiotic with very interesting broader applications.


Quantum Computing

Rapid Scale-Up of Commercial Ion-Trap Quantum Computers – IEEE Spectrum

CCNY team in quantum algorithm breakthrough

DEvelopments in terms of hardware and software for Quntum computing are nowadays running very fast every week. The proposed quantum algorithm above, opens a new venue to use the new quantum devices to study problems which are quite almost impossible to study beyond a given complexity, using classical computers.




Catalyzing a zero-carbon world by harvesting energy from living cells

The molecular catalyst can be used to make zero-emission materials such as plastics and polymers from biomass-based wasteful feedstock.


Nuscale Will Get 25% More From Modular Nuclear Reactors For Breakthrough Economic Value |

Nuscale modules will be reportedly superior to all of the conventional large nuclear reactors and competitive against many natural gas and clean coal alternatives. Small nuclear plant modules are very safe, cost efficient and portable making them the most interesting energy solution that is not discussed at the moment, as much as it should.


Burning Iron for Fuel: Metal Powder as Renewable Energy

The iron acts as a kind of clean battery for combustion processes, a simple concept that is getting adopted very slowly right now, before since fossil fuels were so competitive for centuries. Iron powder production would need to be scaled up for wider energy usage, however it is more superior than lithium batteries, for instance.


Machine Learning

Google AI Blog: Mitigating Unfair Bias in ML Models with the MinDiff Framework

MinDiff is a new regularization technique for effectively and efficiently mitigating unfair biases when training ML models being used as classifiers that tend to exhibit such anomalies towards certain social subgroups.


Facebook is using AI to stem dangerous and false posts

The system is about combining AI and human reviewers to make less total mistakes, and by definition imperfect, however it would not be so strongly criticized if it was open source and transparent for all parties.


Machine learning for making machines: Applying visual search to mechanical parts

Researchers develop machine-learning optimizer to slash product design costs

Mechanical Engineering has been revolutionized by software development since the 60s. New AI tools can now be used for design optimization for a general product, minimizing product design costs, time and human resources in terms of engineers needed.


HAMLET: A platform to simplify AI research and development

The envisaged platform allows researchers to browse through existing machine learning models and train or evaluate their own algorithms, thus aiding their research and development efforts and democratizing machine learning models developed around the world.