2024: The year in which the future dawns

0

Days of the future have passed…

There are films that I enjoyed growing up that are heavily dated with themselves and present visions of a possible future on dates that are now in the past. The ultimate example is Stanley Kubricks’2001: A Space Odyssey‘, depicting the arrival of commercial passenger space and intelligent machines 21 years ago. Other examples include the beautifully executed dystopian vision of Ridley Scotts’Bladerunner‘ (set in Los Angeles in November 2019), the flying DeLorean and ‘Mr. fusion generator from ‘Back to the Future: Part II‘ (which takes us “forward” to the Hill Valley of 2015) and the war with the machines in ‘The Terminator‘ and ‘Terminator 2: Judgment Day‘ (where Skynet was brought online and started a nuclear war in 1997).

The technologies showcased in these sci-fi touchstones may not have materialized by the announced dates, but we’re getting very close – and a number of developments suggest that 2024 could be the year in which reality does catching up with those sci-fi futures.

Coming soon to a reality near you…

If we look at the future technologies around which the plots of these films are built, some are within reach. The prevailing view among futurists is that we tend to overestimate the progress that will be made in two years and underestimate the progress that will be made in ten years. If you fall squarely into this trap, the following developments are on the horizon and could be with us within just 24 months…

Fusion Power: Whether it avoids the need to use a bolt of lightning to provide the 1.21 GW of power needed for time travel, or provides the impetus to visit a monolith in orbit around Jupiter, fusion is the science fiction power source of choice. In today’s carbon-conscious world, Fusion delivers the promise of secure, carbon-neutral, and virtually unlimited power supply. So far, fusion reactors are not yet required to produce more electricity than is input, which means they don’t actually produce electricity. The goal is to reach and exceed the break-even point. By far the largest fusion energy project, ITER (the multinational tokamak project being built in France), is targeting December 2025 for the first plasma and predicts that it will eventually have a 10-fold power input yield (or Q = 10 in industrial terms) will reach ). A number of fusion startups are struggling to beat ITER to break even. In March 2022, British fusion start-up Tokamak Energy reached a milestone of 100 million degrees Celsius in its ST40 experimental reactor. Tokamak Energy had also maintained plasma in a previous reactor for more than 24 hours. This paves the way for upgrades to ST40 that could potentially allow the reactor to run for an extended period and break even before the end of 2024…

Commercial space: PanAm may not be trading anymore, but the idea of ​​taking a commercial flight to a space station (all to the tune of the waltz “The Blue Danube”) is an enduring image. Some may argue that the era of commercial spaceflight began with paid multi-millionaire tourist flights aboard government-operated aircraft or, more recently, SpaceX’s Falcon 9/Crew Dragon. Others might cite the suborbital exploits of certain high-profile billionaires that became widely publicized in 2021. However, the promise of dramatically lower cost per kilogram to orbit that will come with fully reusable spacecraft has yet to be fulfilled. However, with SpaceX expected to make the first orbital flights of its two-stage, fully and rapidly reusable Starship/Super Heavy rocket later this year, this system could be crew-qualified and even deployed as part of humanity’s return to the moon by the end 2024 …

Sentient Machines: Given how often sentient machines are portrayed as turning against their creators, humanity’s fascination with them may reveal a deeply masochistic streak in our species. For every gun-wielding T-800, bone-crunching Nexus-6 Replicant, or chillingly murderous HAL-9000, there’s a friendly and supportive Artoo unit or Pinocchio-esque Commander Data. Great strides have been made in artificial intelligence in recent years, with AI models scaling from millions to billions of parameters tuned via the learning process. We are reaching a tipping point… our AI models may shortly exceed the parametric capacity of the human brain. A wide range of specialized AI hardware has been developed in parallel, with different companies taking different approaches.

With current AI techniques based on neural networks and relying heavily on massively parallel matrix multiplication, speeding up these “matmul” operations is a focus area. Companies like Lightmatter are exploring a shift from electronics to silicon to photonics, computing with light. This has the potential to enable very fast processing without the heat dissipation issues associated with silicon chips, and Lightmatter claims its photonic processors achieve 7x the computational density than the “benchmark” DGX-A100 AI accelerators.

To stay with silicon, various companies are looking at ways to provide lower power solutions that will enable AI in a much wider range of devices. Two notable examples are Brainchip’s Akida processors and Mythic’s analogue (or “analogue” in British English) AI processors. Brainchip’s Akida takes a “neuromorphic” approach, mimicking the spines of neurons in the human brain. In contrast, Mythic’s chips leverage the electronically simpler designs for multiply and add operations in the analog domain for a different approach to low-power AI inference solutions.

However, none of these solutions promise anything that comes close to sentience. We have to look elsewhere for that. After recently announcing the “bow” wafer-on-wafer upgrade of its already amazing Colossus Mk II Intelligence Processing Unit (or IPU), Graphcore has said it is working on a £100m AI supercomputer that it will called “Good Computer”. Named after computer pioneer Jack Good (who was consulted by Stanley Kubrick, among others, on the design and rendering of HAL-9000), this machine aims squarely to run AIs that match and exceed the parametric capacity of the human brain. The good computer will bring AI models from the range of around 100 billion parameters today to 100 trillion parameters or more. If our sentience and consciousness are emergent properties of the sheer neurological complexity of our brain, this machine could be among the first to show anything approaching recognizable consciousness. When is it expected to be completed and online? You guessed it, 2024…

Leaving the law behind…

With science fiction breakthroughs emerging in many areas, laws and regulations need to be reconsidered in light of these new developments.

Some will be straightforward but necessary extensions to existing rules. For power generation, aerospace, developments in fusion power and commercial space companies are an extension of existing regulatory regimes. With frameworks that already allow for fission energy and space launches, it’s not too difficult for the law to keep up with technology in those areas.

In the field of AI, on the other hand, technology is always ahead of regulation. As lawmakers grapple with rules designed to address the potential ethical, trustworthy, and regulatory challenges of today’s AI, tomorrow’s AI will offer a different and much darker perspective. Circulating in the EU is the draft AI regulation, which addresses the protections needed for humans when interacting with AI systems, specifically the various prohibited or high-risk AI use cases that the draft regulation defines. Other jurisdictions take a more piecemeal approach, opting instead to regulate for specific use cases. An example of this more domain-centric approach are the controls in place in New York around the use of AI and algorithmic processes in hiring and recruitment.

History teaches us that it takes time to debate and pass laws, and often longer to enforce them. While no firm deadlines have been set for the EU AI Regulation, it is expected that the legislative process itself could take another year and in general the EU tends to give an additional 18 months to 2 years after legislation is enacted before before they come into effect. On this timeline, the EU AI Regulation itself might not come into force before the end of 2024…

Even if it were, does it, or any of the currently proposed AI legislation, go far enough to recognize and protect all relevant parties? Laws under discussion are generally framed around the protection of people as an unspoken axiomatic assumption. This assumption may need to be reconsidered. If there is a plausible perspective for the emergence of machine consciousness in the medium term, shouldn’t we also think about what legal recognition and protection these intelligent machines might need?

Have you tried turning it off and on again…?

One need only look at the treatment of what we currently refer to as “AIs” in a range of contexts to see the abuses that could be made of intelligences denied “human rights”. Whether we’re being rude to voice assistants or murdering non-playable characters in video games, we’re not used to having sympathy for the machine. This is perhaps hardly surprising given that in many places a lack of public empathy allows monstrous circumstances to befall people we know to have similar intellectual and emotional responses to ourselves, with little or no protection. What hope, then, is there of kindness or concern for anything so different from us?

For the earliest machine intelligences with even a glimmer of consciousness, hopefully a spirit of parental care will protect them from the researchers who work with them. However, before commercialization strays too far in the direction of exploitation, it is important that we as a society consider what legal protections the children of our minds should be given.

After all, resetting a conscious machine at a certain point could amount to murder.

The next 24 months…

As sci-fi author William Gibson famously said, “The future is already here, it’s just not very evenly distributed”. The companies that have the best chance of succeeding when sci-fi becomes a reality are the ones that face the future early.

Share.

Comments are closed.