Top 10 tech trends in 2018 — Issue #3

Top 10 tech trends in 2018

2018 is off to a great start with the launch of world’s strongest rocket, SpaceX Falcon Heavy with Tesla Roadster as a payload into space. It’s a technological miracle, and also a stepping stone to build even more powerful reusable rockets to send humans to mars.

Okay, you ask how does someone predict future. Here’s a simple quote that answers your question:

The best way to predict the future is to create it.

You guessed it right, that’s what Elon Musk does with:

1. Tesla (Electric Cars)
2. SpaceX (Reusable Rockets)
3. OpenAI (Artificial General Intelligence)
4. Boring Co (Solving Road Congestion)
5. Neuralink (Brain Computer Interface).

But, what’s the easiest way to predict the future?

The easiest way to predict the future is to look for people who are creating it now.

Apply that principle and you will find trends!

(Spoiler alert) Almost all the trends you will see in the future will be powered by deep learning technology ;)

These are the top 10 tech trends in 2018:

#1 Autonomous Driving

2018 is going to be a great year of autonomous driving industry.

An autonomous car and unmanned ground vehicle is a vehicle that is capable of sensing its environment and navigating without human input — Wikipedia

In the past few years, Autonomous driving technology went from “how can you imagine?” to “it’s a toy” to “very possible” to “all auto manufacturers announcing their plans for self-driving cars”. That’s how technology will impact individuals and companies.

If you have used driver assistance, lane control assistance or remove controlled parking assistance then you have experienced rudimentary levels of autonomous driving.

In fact, there are exactly five levels of autonomy driving. First level is driver assistance; second level provides semi-autonomous driving assistance systems like lane control assistance; third level is highly automated driving which only works under certain conditions like highway; fourth level is fully autonomous driving, where your car handle most complex driving situations; fifth level autonomy (aka Full Automation) means, the vehicle can completely drive itself in all scenarios without any human intervention. In fact, you don’t need a driver’s license at that point.

Autonomous driving technology is powerful and impactful because it can potentially save millions of human lives from accidents, also save 250+ million hours of commuting time per year. It makes cities save space on parking because the vehicles can drive themselves, moving someone or something from one place to another all the time.

Autonomous driving market is so huge (~$83 Billion by 2025, ~3 Trillion by 2035) that every car manufacturer, ride hailing company and big technology companies wants in. You can see the list of companies who are invested in this space here.

Autonomous driving technology uses combinations of LiDAR, Radar and cameras to sense and track objects around the vehicle to navigate. Under the hood, you have deep learning algorithms running on GPUs installed in the vehicle, using the data collected from those sensors to analyze and make decisions.

#2 Augmented Reality

Augmented Reality is a technology that superimposes a computer-generated image on a user’s view of the real world, thus providing a composite view.

Augmented Reality (AR) is one of the hottest areas in tech industry and it’s already ready for 100s of millions of consumers to use and experience. Snapchat Lenses, Pokemon Go were the first ones to introduce working Augmented Reality to the world.

Last September, Apple announced its plans to bring AR to the world through the release of ARKit, a software development kit for iOS developers. After Apple’s announcement, Google announced similar plans to bring AR into Android devices through ARCore dev kit. This enables developers all around the world, to bring new experiences based on AR to everyone.

ARKit blends digital objects and information with the environment around you, taking apps far beyond the screen and freeing them to interact with the real world in entirely new ways — Apple

AR needs accurate mapping of your surroundings to place digital objects. Visual Inertial Odometry (VIO) technology enables the application to accurately track the world around you. VIO fuses live camera data with motion data (accelerometer, gyroscope etc.,) to sense how the user moves within a room with a high degree of accuracy. All of this is only possible because of advancement in computer vision and availability of high performance mobile CPU and GPUs to analyze and render 3D digital objects onto your mobile phone in real time.

While mobile AR is great for consumers and developers, it can’t provide persistent, collaborative and immersive experiences yet. Companies like Magic Leap, Microsoft and MetaVision are building special hardware and holographic technology to power immersive and real-world blending experiences through head mounted devices.

If you want to know how we compute in the future, check Magic Leap One, a creator edition from the most hyped AR company, Magic Leap.

#3 Robo Advisors

Robo-advisors or Robo-advisers are a class of financial adviser that provide financial advice or Investment management online with moderate to minimal human intervention — Wikipedia

If you are using Wealthfront or Betterment then you already know or using robo-advisor(s) . Robo-advisors help you handle sophisticated tasks, such as tax-loss harvesting, investment selection and retirement planning. Robo-advisors is a huge industry with an explosive growth. It is estimated that client assets managed by robo-advisors hit $60 billion at year-end 2015, and is projected to reach $2 trillion by year 2020.

Unlike human advisors, robo-advisors are scalable, available 24/7 and service every consumer who wants to optimize and invest better. Robo-advisors can analyze trends, data and news about your portfolio in milliseconds using AI technology.

With robo-advisors, you can configure risk appetite for your investment and relax while your portfolio turns green 🤑🤑🤑

#4 AI doctor

Due to recent advancements in deep learning and computer vision, AI software is able to detect and diagnose many kinds of diseases/illness better than best human doctors.

At the time of writing this post, Alphabet’s Verily announced that they have developed a new AI algorithm which can accurately deduce an individual’s age, blood pressure, and whether or not they smoke just by looking at scans of patient’s eye. Doctors may potentially use that data to analyze a patient’s cardiovascular risk without needing a blood test.

You should think AI doctors as personal AI assistants to doctors, helping them save time, analyze faster, make less mistakes while diagnosing a patient. For example, Arterys is the first FDA approved AI cloud software to be used in a clinical setting, for liver, lung and Cardiac MR image analysis.

A number of private and public companies are working very hard to solve health using AI and make it accessible to everyone with mobile device with a camera. Medical futurist did good job listing most of tech companies working on AI health.

If you are curious about this space, checkout recent AI health report from JASON advisory group (elite scientists who advises U.S government).

#5 Voice Editing

Goal of any technology should be making things cheap, efficient and accessible to everyone. Voice-editing is one such technology. You can think voice-editing as “Photoshop-for-Voice”. It will allow you to instantly alter voice recording to include words or phrases the original speaker never uttered, in what sounds like their voice.

We are on the verge on creating computer generated audio completely indistinguishable from a real human’s voice. This has huge implications to the society, journalism, authenticity of an audio evidence presented in a court, voice IDs and lot more.

This technology, powered by deep generative neural networks (see this quora post to understand what it is in simple terms) saves lots of human hours spent on correcting and redoing the audio recordings for voice-over artists, podcast creators and audio book recorders.

Products with Voice-editing technology: Lyrebird (in beta), Deepmind’s WaveNet (not accessible to everyone) and Adobe’s Voco (potential release in 2018).

#6 Photo Synthesis

This is not the photosynthesis we all know (i.e., a process used by plants and other organisms to convert light energy into chemical energy).

Photo synthesis we discuss here artificially generates high-quality images/photos from a text description. You can think of this as Google Image search, except image shown to you is not a “real” image captured by a camera but one that looks real enough for your eyes.

It is powered by generative AI algorithms called GANs or Generative Adversarial Networks. GANs are used for generating “fake-but-realistic” data from real data. Facebook’s AI research director Yann LeCun called GANs — “the most interesting idea in the last 10 years in ML.”

Here’s a realistic fake image generated by GANs from a text description:

A flower with overlapping pink petals surrounding a ring of yellow filaments

Retail industry spends huge amounts of money on photo shoots for each new run of clothing. This involves hiring a professional photographer, models, and a studio. With this technology, retailers can automatically generate images of the garment on a person of any size, shape, or wearing any kind of shoes, given only a picture of the garment laid out on a plain surface. Vue.ai is an Indian AI company backed by Sequoia already selling this technology to retailers like Macy’s and Levis. If you are into modelling business, you should worry about your job!

#7 Fake videos & Propabots

By now, you must have realized the power of generative learning (aka Generative Adversarial Networks). Now, using same foundational technology, we can generate realistic fake videos and images. And this is just the beginning of “fake news” and you should worry.

Unsupervised Image-to-Image Translation Networks from Nvidia

Here’s a fake video of obama generated by AI

Technology is scary. You’re going to have negative consequences. But the positive potential far outweighs the bad.

For example, a company called Pinscreen has created a technology (similar to iPhone X Animoji) to generate your face in 3D from a single still selfie image. The same technology can be used in CGI (Computer Generated Imagery) to slash huge animation and production costs.

#8 AI building AI

Tech companies like Google, Facebook and OpenAI are investing and betting heavily on a technology called AutoML because of shortage in AI experts (only 10,000 available to hire).

This technology enables non-tech companies and individuals without AI experience and/or background build machine learning models easily (see AutoML from Google Cloud) without much help from AI scientists. Almost all products we use everyday is powered by Machine Learning(ML): Google’s Search, Facebook’s Face Detection, Apple’s Siri etc., Machine learning models behind those products are built by top AI scientists. Most of current machine learning models are based on Neural Network architecture. AI building AI (or learning to learn) technology runs a neural network algorithm to scan for best neural network to build a custom model that works for your data.

This is how AI technology will be democratized.

#9 Decentralized Crypto Exchanges

Crypto exchanges allow you trade one cryptocurrencies with the another. Almost all crypto exchanges that exist today are centralized. Centralized exchange provide liquidity, maintain order-books, enforce AML (anti-money laundering) and KYC (know your customer) rules on customers depending on jurisdiction.

Centralized exchanges are easy to access and use, and provide advanced tools and trading functionalities such as margin trading, stop loss and lending.

However, centralized exchanges pose a risk for your funds because you don’t own the private key of your funds/wallets. People have lost millions worth of their cryptocurrencies like bitcoin due to hacking attacks. Just last month, hackers hacked 500 million dollars worth of cryptocurrencies from Coincheck exchange (Japan).

Welcome to Decentralized exchanges, these are exchanges with no central authority, so there is no censorship (i.e., nobody can freeze your funds) and are trustless by nature (you stay control of your money). Decentralized exchanges allow users to trade peer-to-peer (P2P), so those trades occur directly between users via bitcoin-like blockchain, so anyone can lookup and validate a trade on blockchain.

Decentralized exchanges are the future of crypto trading. 0xproject, OmiseGo, AirSwap are some building it.

#10 Private Coins

“Bitcoin is anonymous”, “Bitcoin is only used by criminals and for illegal activities” are myths. Bitcoin offers same level of privacy that you get if you buy something using cash in a store where no one recognizes you. In this scenario, your transactions are anonymous, but the transaction amounts can easily be traced. And if someone is able to determine your identity, all privacy quickly evaporates. So Bitcoin is pseudo-anonymous in nature.

All Bitcoin transactions are stored publicly and permanently on the network, which means anyone can see the balance and transactions of any Bitcoin address. However, the identity of the user behind an address remains unknown until information is revealed during a purchase or in other circumstances. This is one reason why Bitcoin addresses should only be used once. Always remember that it is your responsibility to adopt good practices in order to protect your privacy — bitcoin.org

If you are a business owner and thinking of using Bitcoin/Litecoin as method of payment, you should reconsider because your competitor or literally anyone could see your account balance and transactions using your Bitcoin public address on Bitcoin’s public ledger.

Stanford’s Applied Cryptography Group is working on a new proposal called Bulletproofs to make Bitcoin transactions anonymous and confidential. It is going to take a while for researchers and bitcoin developers to implement this technology on Bitcoin. Bitcoin community may take even longer to reach consensus without a hard-fork (if there is chance of hard fork at that time, it would be popcorn worthy to watch the Bitcoin upgrade). Hard fork happens when majority of miners in Bitcoin don’t agree to protocol upgrade (a la Bitcoin Cash).

Private/Anonymous coins are cryptocurrencies that have public ledgers, but transaction information is obfuscated in varying degrees to protect the privacy of users.

You may heard that some websites are using user’s computers to mine cryptocurrency called Monero. At the time of this writing, it was discovered that a hacker injected a crypto script (aka cryptojacking) into LA times website to mine Monero.

Monero is one of the most popular private coins.

Monero uses Ring Signatures to make it difficult to trace the parties involved in a transaction because transaction signatures are shared by a large group of people; as a result, associating specific users with a transaction is very difficult. It also provides additional privacy by obscuring transaction amounts.

Zcash is relatively new (created in 2016) cryptocurrency that can fully protect the privacy of transactions using zero-knowledge cryptography.

Because of the untraceability property, no one can blacklist units of currency due to their association in previous transactions. One downside to it: If an exchange gets hacked, and hacker steals private cryptocurrencies, there is nothing that can be done to stop hacker from selling or transferring.

Private coins are already receiving adoption from Musicians, Shopify, WooCommerce etc.,

This blog post turned out to be longer than I expected. If you are reading this line, my mission is accomplished! Stay excited for the future ;)

Originally published on The Tech Fusion

Top 10 tech trends in 2018 — Issue #3 was originally published in Hacker Noon on Medium, where people are continuing the conversation by highlighting and responding to this story.