Nothing vast enters the life of mortals without a curse – Sophocles, and the intro to Netflix, The Social Dilemma.
I was really excited to see that Facebook’s Ron Ho was presenting a keynote at IMAPS Symposium last month – not because of the topic he was going to discuss, but because it opened up the opportunity for me to talk about a documentary-drama that I recently watched called The Social Dilemma. If you haven’t heard of it or seen it, I highly recommend you put it on your Netflix watch list right now… especially if you have teenagers at home.
Based on interviews with the technology experts who created Google, Facebook, Twitter, Pinterest, Instagram, Snapchat, and others, this film reveals the very darkest side of social media, and the negative impact it’s having on society as a whole. It is impacting our mental health, fueling bullies, being weaponized to polarize nations, perpetuating ‘fake news’, the list goes on.
Watching it made me question my role in promoting heterogeneous integration because its advancements will provide the backbone to artificial intelligence (AI) and machine learning (ML). And the more advanced these technologies become, the more harm it can cause when put in the wrong hands. And sometimes, the wrong hands masquerade as a company professing to make the world a better place. How can I play a part in that? Luckily, I came to my senses and decided instead to use my voice to raise awareness instead.
They Sell our Data – What’s the Big Deal?
I already knew that when I search Google or click on a Facebook post, enter in a GPS location, or make a purchase on Amazon, AI engines are at work learning my locations, studying my reading and buying habits, all under the guise of making my life easier. I knew that the personal data gleaned by these platforms is the currency they trade for cash. A lesson I learned 30 years ago in a mass communications class rings truer than ever: If it’s free, YOU are the product. Except that in this case, it’s not always free. We pay for our Prime Memberships, and our Internet connection, and our devices. But I digress…
And while I’m also was aware of the addictive nature of social media platforms and the impact that is having on society, I didn’t fully understand, until watching The Social Dilemma, that not only are these organizations following and tracking our every move, there are “teams of engineers working behind the scenes whose job it is to use our psychology against us.” In addition to selling our information to the highest bidder, they are manipulating what content we are served to either perpetuate or change our ways of thinking. Nowhere has this been more evident than during the recent US election. And that’s really just the tip of the iceberg. Really – watch the film. It’s 90 minutes well spent.
Note that it is not lost on me that I am an active user of social media platforms, including Facebook, Instagram, LinkedIn, and Twitter, for both personal and business purposes. Does that make me a hypocrite? I don’t think so: If anything it casts me as an indignant victim. If I was unaware of the depths of social media manipulation, what about all the innocent users who don’t work in the technology realm?
So, by now you’re asking, what does all this have to do with Ron Ho’s keynote at IMAPS? Read on.
Better Living Through VR Glasses
The reason Ron Ho was invited to speak at the IMAPS Symposium is that two years ago, he was hired to be Facebook’s director of Silicon Engineering in hopes of finally doing something with the virtual reality (VR) capabilities in its portfolio thanks to the 2014 acquisition of Oculus. According to this 2019 podcast, that investment has yet to allow Facebook to realize its vision of VR as the future of engagement.
The way Ho spins the story, Facebook thrives on shared connections to drive engagement. Its purpose, he says, is connecting people. By developing VR glasses that aren’t the clunky, nerdy goggles of today’s VR gaming, but rather are as unobtrusive as those we wear for correcting vision, Facebook will provide the tools to underpin personal connection. From enhancing virtual meetings and telemedicine visits to translating a menu or what your waiter at a restaurant in Paris is saying, these glasses will be better than a personal assistant.
“Glasses are interesting… they maintain the connection to where you are today. They are vehicles for transmitting bits of contextualized information that improves your day, and improves the ability to connect,” says Ho.
These glasses will not only enhance our work, says Ho, but our life as well. We can share transmitted images with friends, make healthier food choices, and use IR sensing to understand why a baby is crying. Beamforming microphones and speakers improve our conversations.
That awkward moment you experience when you run into someone but can’t place them? Not a problem. These glasses will use facial recognition to identify the person and search your Facebook friend list and provide you with the name just when you need it. No more embarrassing moments.
To pack all this functionality (sensing, edge computing, memory, etc.), Facebook needs heterogeneous integration, explained Ho. Technologies are constrained by volumetric size to fit the space created by interstitial gaps between the frame and the lens.
What’s Under the Hood?
Is it even possible to put all this functionality into glasses? Ho asked. He called for improvements across the board:
- Silicon needs more performance relative to power
- It needs to fit inside the form factor of a pair of glasses
- Thermal solutions must ensure heat dissipation around the skin
- It needs high bandwidth wireless connectivity
- Compute power needs to be increased while maintaining low latency
- Graphics and display must be transparent
- Localized compute is needed to reduce energy consumption and eliminate the cost of data transfer
- AI systems will need embedded neuro-network accelerators
- It needs 2.5D and 3D solutions that fuse sensors and compute, and graphics pipeline paired with display to refresh images at a fast rate, motion sensors so that when the wearer moves their head, the display stays locked to the world.
- New assembly methods are needed
- The ability to stack unique sizes and shapes
- Material diversity that handles thermal cycling and aging
Ho’s job, therefore, is not trivial. He says the journey is only 1% finished. “Silicon has only been alive at Facebook for two years. We’re new to the game,” he said, adding that they are still learning where they have to take Silicon, packaging, and co-design software.
The message to the audience: there is an opportunity to be had here, my friends.
Back to the Social Dilemma
Before I watched the Social Dilemma, I would not have fully questioned Facebook’s motivation for developing these VR glasses. I would have stopped here – with the message about the opportunity. But I did watch it – and it makes me question everything.
On the one hand, Ho offered some compelling and life-improving applications for glasses. But then, this is FACEBOOK we’re talking about. Since when is it a device company? Data is its currency. Any devices it builds serve only to allow it to gather more data and use it to manipulate our behavior.
Remember the bomb that was Google Glasses? People thought they were downright creepy and intrusive. What makes these Facebook glasses any better?
Ho answered questions about the privacy and security of the glasses. He said a set of operating principles will be designed in; real controls that make sense and are intuitive for the user. He also said to actionable controls will prevent data from being sent just anywhere.
But what about the unintended consequences? Are they really unintended? How about that facial recognition software that scans the approaching person and searches for a match on Facebook? Whose privacy is that violating? What if that person doesn’t want to be “identified”? How is this any different than China using WeChat and facial recognition to create profiles and track its citizens? What makes Facebook think people will buy into this?
During a segment of the International Semiconductor Executive Summit Webinar series entitled “The AI Chip Landscape Overview,” Keith Strier, VP Worldwide AI Initiatives, Nvidia noted that high-performance computing (HPC) and AI have converged such that engines of scientific discovery are now engines for economic growth. This is a big turning point in the world, he said. We need to think carefully about the implications this has for society as every nation can now be an “AI nation.” Strier says he spends much of his time working with countries all over the world to develop successful AI programs. He says we have a responsibility to ask questions.
“AI is dual purpose,” he said. “It can be applied towards nefarious purposes. You have to have the right stakeholders not driven by the profit motive.” He added that we need to enact policies and regulations where it makes sense to balance innovation with regulation. We are at a time in history where we see possibilities to transform the world with AI.
“We can’t stop the progress, but we have to have an equally robust narrative and dialogue to make sure the train doesn’t run off the rails.” The healthy development of an AI system requires good leadership and a strong AI policy.
The semiconductor industry is responsible for providing the world with vast capabilities beyond our imagination. We need to understand that the technologies we develop with the intention to save the world, can also destroy it. And sometimes the villains are dressed like heroes. The onus is on each one of us to understand what’s at stake and proceed carefully. So as a supplier, before you see dollar signs in supporting Facebook’s efforts, watch the Social Dilemma, and decide for yourself if it’s worth it. That’s all I ask. ~ FvT