Tech-A-Week: DeepFake – Game of Thrones in real life

Game of Thrones – Faceless Men

For Game of thrones fans the episode when faceless men was introduced was a shocking but confusing one. Faceless men are the ones that have the ability to change their faces and capability to shape shift to appear as a different person. They follow the “many-faced God”, are trained to relinquish their former identifies, feelings, aspirations to become “no-one”. They use their ability to shape shift to efficiently carry out assassinations and carry out their tasks with ease.

In Real life, at least in the digital world, a combination of advancements in machine learning algorithms, faster processing power, huge amount of face data makes it possible to “face-shift” someone with another. This new technology is called “DeepFake”. There are plenty of examples of this new crazy technology online.

DeepFake: Layman’s definition

DeepFake is fancy name for “Face swapping” in videos. Take a famous person’s video, or a movie scene or any video for that matter then using software to replace Person A’s face with Person B’s face! .

Why is this a big deal?

Changing a person’s face in a photo is not that difficult, anybody with a decent phone or computer photo editing software can do it. Photos are stationary and doesn’t move, just replacing a face is not a big deal. However, video is a whole another ball game. you need a sophisticated and powerful software to replace a face in a video. Hollywood studios could do it with their arsenal of computer graphics machines and technical experts. check out the trailer and the backstory for “Gemini Man” movie.

So far it was not available for common people, but now a technical advancement called “DeepFake” is enabling anybody with a phone to create totally compelling video by replacing someones face. On one hand it is a great, exciting and on the other it is disturbing someone could do this with ease just with a phone.

Pretty impressive Bill Hader video in which his face changes to Tom Cruise and to Arnold Schwarzenegger.

Advancements in DeepFake:

In the last few months DeepFake “how to” videos and libraries have been exploding online. Mostly they are for entertaintment value. Some are complex step-by-step instructions (check the section towards the end of this blog if you are technically inclined to try this yourself).

There are also simple, user friendly apps in phones available to “wow” your friends (and probably creep out some).

One such example is an app called Reface ( Android and iOS) from an Ukranian startup, that is taking the online world by storm.

Experts are already cautioning against the negative consequences of this technology. It will be really difficult to distinguish (at least online) between real and fake videos. What will be negative consequence of the app especially in the election year?

Interesto Meter

Investment Potential: Low

Entertainment value: Huge

Social Impact: Huge (potential to wreck havoc with a deluge of fake videos created by rival political ideologues, fake videos by disgruntled employees or ex-lovers etc)

Technically inclined!?: Continue reading…

Face or image recognition is not rocket science these days. Facebook and Google have been using image recognition in their platforms for several years now. If you ever used Google photos you’ll have experienced image recognition. Google photos can accurately identify a “Cat” without needing to supply “meta-data” to specify a photo contains contain. It recognizes by training itself using millions of cat photos and pin-pointedly identify Cat and even has the ability to distinguish between Cat and for example, Tiger.

However, recently there are advancements in the science of video editing to replace a face of pone person with another has been making big strides. below are two open source software examples that provides enough arsenals for someone to create DeepFake programs with relatives little effort.

  1. Faceswap
  2. DeepFace Lab

Here is a video that gives step by step instructions using DeepFaceLab to reface Elon Musk to use Robert Downing Jr’s face.

Last Updated on September 8, 2020 by SK

This week’s Newsreel

Last Updated on September 6, 2020 by SK

What’s your management style?

In East Indian mythology there are two great epics. One is called Ramayan and other is called Mahabharata. The central story of both these revolves around victory of good over evil. In one story  Lord Rama leads his army to defeat Ravana (demon king) in Sri Lanka. While in the second Lord Krishna oversees Pandavas (the good guys) defeat Kauravas (The villains) in the battle of Kurukshektra.

Lead from the front

In Ramayan,  Rama is the best warrior of his side. He leads his army from the front, creates strategies and directs a diverse army of monkeys and people to execute his plan to meet the objective of rescuing his wife from the demon king. His people are happy to follow orders and want to get all the appreciation for being the best executors. Rama sets direction and also tells people what to do during difficult times.

Ultimately they won the war and the final outcome was achieved.

Guide from behind

On the other hand Krishna was clear with Arjuna (The good prince), I won’t fight the battle, I won’t pick up any weapon, I would only be there on your chariot as a charioteer. And he did what he said. He never picked up the weapon and he never fought.

Still, Pandavas won the war and the final outcome was achieved.

Two different styles

So, what was different? It was their managerial style and it was also the type of people who were being lead.

Rama was leading an army of primates and ordinary people who were not skilled fighters and they were looking for direction.

While on other hand, Krishna was leading Arjuna who was one of the best archer of his time. While  Rama’s role was to show it and lead from the front, Krishna played the role of a coach whose job was to remove cobwebs from his protege’s mind. Krishna couldn’t teach Arjuna archery but he could definitely help him see things from a very different perspective.

Basic differences in two styles:

 Rama: A skilled warrior, lead monkeys, was emotional, gave precise roles and instructions, motivated the army to fight for his cause

Krishna: A calm and collected coach, works with the best professionals, provides strategic clarity, allows team members to take the lead, fights for the cause of the team, does not depict his true emotions

What’s your style?

Look at your team and reflect what type of leader you are? One who keeps answering/solving problems for the team or who asks relevant questions from their team so that they can find their own solution.

Are you someone who commands/directs all the time  or someone who clarifies doubts and allows their team to find their own ways.

Are you someone who has unskilled people in the team for the given job or you have the brightest experts in their area getting stuck with issues? 

A capable team doesn’t want you to micro-manage or show how things are done, they want to know the meaning of their task and how it makes a difference in this world.

They are Arjuna’s who don’t necessarily seek more skill/knowledge but they need someone to clarify the cobwebs in their mind. If you apply Rama’s style on them, you are bound to fail as a manager.

On the other hand if there are people who aren’t skilled enough but rely on your expertise to sail through, then  Rama’s style is appropriate.

Who are you?

It is good for you to reflect and think what managerial style will bring the best result for you and your team?

Are you a  Rama or Krishna?

The Managing Leader vs the Coaching Leader!

Disclaimer: Received the crux of the idea from a WhatsApp group forward.

Last Updated on September 3, 2020 by SK

My first blog with BlueData

Checkout the blog here:

Prior to joining BlueData recently, I worked at Cloudera (through the acquisition of — and prior to that at Pivotal (via EMC / Greenplum) where I led product management for their Hadoop software business. Back in 2012, I still remember my first demo and presentation about Hadoop to a Fortune 500 company: I covered the basics of Hadoop and the type of problems it can solve for enterprises. A few years later, I was doing similar presentations focused on increasing Hadoop adoption at other large enterprises. There was less of a need to teach the basics of Hadoop; but the time-consuming complexity of the traditional bare-metal deployment model for Hadoop often stalled adoption. And there were significant challenges to meet the enterprise-grade security and operational requirements typically taken for granted in traditional data warehousing and database systems…(more)

Last Updated on June 16, 2017 by SK

Sept 24, 2016: TeqBiz Today

  • Eleven reasons to be excited about the future of technology. (medium – Chris Dixon)
    • Pretty good list of technologies.
    • Though he missed couple of exciting ones such as:
      • Quantum computing
      • Stem Cells
      • Gene Editing
      • 3D bio-printing.
  • Are Silicon Valley’s Chaos Makers Smart or Insane? (FEE)
  • Oracle has announced that they are bringing the same public cloud infrastructure to enterprise private cloud (Forbes).
    • Meaning, an enterprise can deploy the same tools/hardware that Oracle uses in their public cloud offering. Given an incentive for enterprises to familiarize themselves with the Oracle cloud environment and then move to public cloud whenever they feel comfortable.
    • Smart move on Oracle’s part.
    • Little noticed feature, but one that will have bigger impact on the cloud world in general and for oracle in particular
    • Check out this pdf for more detail.
  • The law of accelerating returns (Ray Kurzweil)
    • Steve Jurvetson called this graph the most important graph ever produced.

Last Updated on September 2, 2020 by SK

Sept 20,2016: TeqBiz Today

  • Here is a link to Intel’s presentation on why FPGA is going to play a major role in datacenter and cloud industries (presentation).
    • Intel pretty much makes a big bet on FPGA + CPU for reducing cost and increasing performance of data centers at much lower footprint. Argument is that, with the powerful FPGA co-processor in a single dye datacenter operators could save on power cost and cooling, and get better performance using a fraction of nodes.
  • Google launches Allo with Google Assitant (new version of Google Now?) (Engadget)
    • May be Google’s answer to Siri?
  • Facebook is augmenting its team that works on the secretive “Building 8” project (Recode)
    • They just acquired a small hardware startup called Nascent Objects that builds “modular electronics platform”.
    • May be a similar project like Google’s suspended “Project Ara”?
  • Oracle is getting its game face on the cloud and it could be a series threat to dominance enjoyed by AWS? (VentureBeat)
    • Enterprises are moving to cloud for sure, it is given. However, the question is what workloads will move to cloud. Large companies are still worried about security and governance issues. They might be willing to move non-critical data and services to cloud and keep their mission critical services running in their own secure data centers, where they’ve much better control and accountability.
    • With rapid advancement in the technology, especially in NoSQL, BigData and Cloud, a  typical enterprise customer is pretty confused about the array of proprietary and open sources technologies. It seems like every month there is a new technology (close or open source) that is trying to do better than the one came out the month before. Its all too confusing for enterprises.
    • Providing a familiar platform, ex. MySQL/SQLServer, some of the concerns could be addressed. Given that, a cloud provider that is familiar with the way large enterprises work, have a presence in the doorstep, provide integration with services on-premise and understand how to decisions are made will win in the long run. Here is where Oracle, Microsoft and IBM have  huge leg up when comparing non-traditional enterprise players (AWS and Google). Amazon and Google have to step up their game to address enterprise market with better messaging and offerings that integrate with existing on-premise platforms and also understand/sympathize with enterprise customer’s needs.

Last Updated on September 2, 2020 by SK