Ep. 6 | Ah! It's time to pay attention to robots and quantum

Quote of the week:

"We see a future where general-purpose robots are as ubiquitous as cars, helping people to do work that needs doing.” - Geordie Rose, Sanctuary AI, the creators of general-purpose robot, Phoenix.

Hi, folks! To me, the past month has been defined by a flurry of developments in robotics and quantum ... so let's talk about that.

First, let me remind you why we should talk about that, and why you should care.

For the past year, I've been making the point that after a decade of digital transformation, institutions must now focus on the next paradigm, and being both strategic and [more importantly] resilient in that landscape.

This is happening. Society is moving to a new 'post-digital' paradigm, and I often use the acronym DREAM-C to cluster the relevant innovative technologies, underpinned by computing, that define this new world.

DREAM-C:

?? Distributed ledger technology (DLT)

?? Robotics (incl. process automation)

?? Extended reality (incl. AR & VR)

?? Artificial intelligence (AI)

?? Mutualisation

?? Computing (cloud & quantum)

(Note, whilst I'm generally talking about post-digital from a financial services perspective, it's a society-level evolution.)

So, although the mainstream finance world has really focused on innovation in DLT and AI, we are right now at the point where some of these other technologies meaningfully come in to play. Last time I mentioned Meta 's innovation in wearables, with Orion, for example. This week, we're going to spend some time with robotics and quantum...

In personal news, I've had a great October thus far! AFME (Association for Financial Markets in Europe) hosted their OPTIC conference and kindly invited me to join a panel to discuss all things AI risk and governance. I spent an awful lot of time at our sponsored coffee cart chatting away about precisely some of the topics that will be covered today, so hopefully it's useful.

Then, I spoke on an ESG panel about the evolution of data and data challenges. This evening, I'm excited to host a dinner of industry execs with Hazelcast ... I'm excited to share takeaways from that in due course.

One bit, two bit, qubit ...

... people are well on the way to taking quantum security seriously. Are you?

?? If you're not really sure what we mean by quantum computing, or want a refresher, I highly recommend this recent article from Moody's that summarises the technology, and the relationship with financial services, really well.

Broadly, there are two perspectives here:

1. quantum computing elevates capability beyond classical computing when it comes to challenging processes and problems (e.g., complex risk modelling and event simulation);
2. from a cybersecurity perspective, quantum computing challenges one of the most fundamentally important features of modern banking: encryption.

Today we don't have a quantum computer available that would be able to break modern encryption protocols, but the algorithms to do so, DO exist. It's just a case of the hardware catching up.

"Experts" are frequently quoted with an estimate for the achievement of this in five to 30 years. Given the exponential growth in technology innovation, I reckon we'll see a major quantum-based public issue in three ... and though (as I'm about to explain) we do have industry standards available, mass adoption will likely take 10 years at least.

How is the industry preparing?

It's easy to get lost in the noise, so I'm going to tell you about just three major developments you should be aware of to have just a general finger-on-the-pulse of how important this is.

?? First, this week at IBM TechXchange 2024, IBM launched Guardium Quantum Safe. It's a platform for assessing cryptographic vulnerabilities in your critical data, and it looks damn cool.

"Gain visibility into your cryptographic posture" is how they're marketing it, and it's focused on data security with regards to both conventional and quantum-enabled risk.

There's a webinar on Thursday 5th December, which you might be interested in joining.

?? Second, working with IBM, NIST launched post-quantum cryptography (PQC) standards a couple of months ago.

Quick overview: there are (generally) two accepted methods for achieving quantum-safety in cryptography. These are:

1. quantum key distributions (QKD)(introduce quantum effects to classic protocols)
2. post-quantum cryptography (PQC)(make classical problems harder with particular algorithms)

The three standards from NIST are all based on PQC, and they are designed to "strengthen modern public-key cryptography infrastructure for the quantum era".

They are FIPS 203 ML-KEM (for general encryption, e.g., accessing secure websites), FIPS 204 ML-DSA (for general-purpose digital signature protocols), and FIPS 205 SLH-DSA (a digital signature scheme).

I've linked the origin for the derived encryption algorithms, for the nerds :-)

?? Third, earlier this year, Europol created the Quantum Safe Financial Forum (QSFF) in collaboration with the European Cybercrime Centre (EC3). As per their own summary:

"The QSFF is a multi-stakeholder effort to address the transition to Post Quantum Cryptography (PQC) across the financial sector, with particular focus in Europe."

If you're reading this thinking "what on earth can my institution do to prepare for this?" the answer is: contribute to the QSFF. They are focused on (1) Planning, (2) Stakeholders and (3) Discovery.

Only last month, Banca d'Italia - Eurosistema hosted a G7 workshop "Building a Quantum Safe Financial System". Major players are taking action on this now, this is not something that becomes relevant in 5-10 years, the opportunities and the challenges exist today, and institutions that move quickly and invest appropriately will save themselves a lot of pain in the future.

What's the deal with "We, Robot"?

... and, actually, what's going on with humanoid robotics more generally?

I'm glad you asked, because today I'm giving you an overview of recent developments in robotics and what to expect looking through to 2025. We are also seeing some early movements when it comes to thinking about what regulation might need to look like (particularly when we talk about 'embodied AI'; that is, AI-powered robots)..

You probably saw Tesla's Optimus robots that were showcased at the "We, Robot" event earlier this month, and did things like dance and serve drinks at a bar. In news that surprised basically nobody given the history of issues with Tesla vehicles, it was kinda a scam. Not only that, the designs were called out by the director of the film I, Robot as being a total rip-off (*gasp!* no way!)

Thus, it was confirmed that human teleoperators were "puppeting" the robots, or as an unnamed engineer confirmed on social media, it was "a human remote assisting".

Although Elon Musk once again disappointed us all, I don't want that to detract from the genuinely incredibly developments in humanoid robotics.

Back in April, the media circuit was abuzz with the idea that we'd all have a home robot by around 2050, largely due to the release of notable humanoid robots like Boston Dynamics' Atlas, and Figure's 01. With the latest developments, I'd put that widespread consumer adoption closer to 2035.

What's going on with the other robots?

?? About 6 months since launching Figure 01, Figure 's robot has got a gig at a BMW manufacturing plant

?? You know about large language models (LLMs), but the more relevant kid in town is the large behaviour model (LBM). That's the focus of the collaboration between Toyota Research Institute (TRI) and Boston Dynamics (the OGs in robotics). I recommend this TechCrunch article all about it.

?? West Japan Railway have introduced a 12-meter-high humanoid robot that's giving Wall-E (and I'm okay with that). It's described as a "multifunctional railway heavy machinery" and requires a single engineer with a VR headset connected to the robot's "eyes". Mechanisation is an important area of innovation, particularly when it comes to reducing workplace accidents and reducing service downtime. The ball is in your court, South West Rail.

?? Neat! We are putting human skin on robots now. Researchers at the University of Tokyo have developed a new technique for binding living human skin to robotic surfaces. 2D and 3D. The quote that stood out to me was that their goal is "to endow robots with the self-healing capabilities inherent in biological skin".

Look, it's not lost on me that every week I identify something new to lose sleep over, but this? This really is the stuff of nightmares. I hate it. It makes me feel uneasy and afraid and like I really should just stop using the internet. That is my sleep paralysis demon. If you want to see the awful images, you can read about it here and if you want to check out the journal entry, that's on Cell and can be found here.

?? Stanford's AI lab introduced 'HumanPlus', which is a new system for humanoids to learn motion and autonomous skills through supervised behaviour cloning and shadowing. They have a really cool video that shows their 6ft robot putting on shoes and walking.

What about regulation?

Robotics is a broad field, and the bit that is particularly interesting and relevant for this discussion is described as "embodied AI".

Now, clearly regulation and standards that apply to AI will also apply to AI systems embedded in robots, so things like the EU AI Act and the Cyber-Resilience Act are relevant.

But with robots, we also have the crossover with hardware legislation, so things like the Product Liability Directive, General Product Safety Regulation (GPSR), the US SAFE Innovation Framework, the UK Automated Vehicles Act, and the European Committee for Electrotechnical Standardisation (CENELEC).

Relevant standards bodies include, as noted, the EU and then the Institute of Electrical and Electronics Engineers (IEEE) and the International Organisation for Standardisation (ISO).

The IEEE Robotics and Automation Society have launched a study group for standards in humanoid robots and I've applied unsuccessfully just 17 times thus far.

? If you're struggling to stay up to date with developments in AI, robotics, and other innovative technologies, let me do the work for you! I also post regularly on the feed. ?

Hope you found this interesting - let me know!