AI dominance, quantum leap and eco challenges ahead :: WRAL.com

Welcome to 2025! It is the time of year for
resolutions and new beginnings. As I’ve done for a few years now, I’ll offer my
technology predictions for the new year in hopes that they are informative and
might inspire you to resolve to make a new beginning by riding the waves of
emerging technology.

In the list below, I’ll offer up what I
believe are the most important emerging technologies to pay attention to, as
they impact our lives and society. I’ll also mention a few that are highly
likely to be big stories with lots of buzz, but in doing so will attempt to
sift the signal from the noise.  Often
the most important tech is not necessarily what gets the most media attention.

I’ll score each technology in terms of its
hype value (1 = under the radar and 5 = in your face) and actual significance
(1 = nice to have and 5 = will change the world in the next 24 months).

Without further ado, let’s get right to it!

AI

HYPE 5
/ SIGNIFICANCE 5

Alphabetical order or not, AI is going to top
every technology list in 2025.  The hype
is through the roof, but for good reason as AI is already changing the
world.  This is too big a topic to count
just one slot on a top 10 list, so I’ll break out a few sub-categories

Small data > big data

HYPE 3 / SIGNIFICANCE 4

The general purpose
transformer (GPT) applied to large language models (LLM) hit its hype peak last
year. LLM’s will continue to grow in daily use, as people shift from using
search engines to AI, but these tools will not look significantly different in 2025
than 2024. The much larger value creation is shifting to use-case specific
transformers, trained on narrow data sets. Some people are beginning to refer
to these as small language models (SLM’s) or focused language models (FLM’s).
Specific depth will create more business value than general breadth.

Agentic AI enters the lexicon

HYPE 4 / SIGNIFICANCE 4

In the FLM space, I
expect the most experimentation to happen in the area of AI agents. In contrast
to traditional LLM’s like ChatGPT which require a human prompt for each
response, AI Agents have autonomy to perceive, decide and act towards specific
goals. The expectation is that we will give AI agents goals to accomplish and
they will be able to do so with minimal human interaction. Think for example of
asking an AI to plan and book your next business trip, perhaps with a few
overall guidelines like a budget limit and your travel style preferences. Done
well, Agentic AI frees our time to focus on higher level tasks, while AI does
more mundane tasks in the background.

User interfaces become more natural

HYPE 2 / SIGNIFICANCE 3

If you’re a college
football fan, you’ve probably already been inundated with advertisements for
Google’s new fully voice-interfaced Gemini AI (or maybe my “smart” TV is just
targeting me). Bringing AI into our daily life workflows means big tech companies
are investing heavily in getting the keyboard out of the way. This is part of a
larger trend called Ambient Invisible Intelligence, where AI is getting
embedded into our lived environment such that we can use it continuously as we
live our lives. Invisible AI is proactive, discrete and context aware, doing
things useful in ways that are congruent with how we already conduct our daily
activities.

Quantum nears practicality

HYPE 2
/ SIGNIFICANCE 2

Quantum computing has seemed stuck in the
world of science fiction for at least two decades. In 2025 we will finally
begin to see it in practical use. IBM made a major breakthrough last March
advancing solutions to decoherence, which is the tendency of quantum bits
(qubits) to be unstable and lose information. This breakthrough greatly reduces
the number of qubits required to maintain result quality so more of the
supercooled computer can be put to practical use. Companies are beginning to
use quantum for real-world applications like complex supply chain logistics and
drug discovery. While quantum advances rarely make the headlines, it is worth
noting that the Triangle could become the global center of excellence, largely
due to IBM Qiskit (the predominantly used sw stack for quantum computing), IonQ
(quantum hardware company with $9B market cap), Duke’s Quantum Center and Atom
Computing ($300M valuation) all based here. I score significance based on
timeline – widespread use is still >24 months away – but rest assured that
the hype score will go through the roof when inevitably quantum computing is
more powerful than any of our cybersecurity and password systems currently in
use. Post-quantum cryptography is a problem we’ll need to solve.

Rise of
the technocracy

HYPE 5
/ SIGNIFICANCE 2

While not an emerging technology itself, it is
worthwhile to talk about expected changes coming with a new administration
populated with many of the richest tech billionaires in the US. Elon Musk,
Peter Thiel and a host of other Trump appointees are supporting the President’s
promise of deregulation and government streamlining. I’m doubtful that all the
promises will amount to much in the near term, outside of increased big
corporate profits. The federal government is a huge ship to steer and changes
will take time to trickle out. There is the possibility that removing
regulations will make it easier to experiment on genetic hacking or other
typically regulatory-constrained fields. But far more likely, deregulation will
lead to a slow-down of technology innovation as companies instead fall back on
old, proven ways of working that put profitability in front of environmental
responsibility, health and safety. Regulations force innovation to thrive,
requiring new solutions to thorny problems. Without – expect more of the same
old same old from market leaders.

The second coming of Web3

HYPE 3
/ SIGNIFICANCE 1

The decentralization movement is a decade old
now and I am already seeing a sharp and loud return of the tech-libertarianism
movement that first fueled cryptocurrency into public discourse circa 2017. Big
banks and brands are leaning into this second wave. It isn’t yet clear if this
is because the market is ready, or simply as a political hedge. The new
administration is absolutely friendly to the Web3 audience and in fairness,
blockchain and distributed ledger technology have come a long way in the last
several years. The renewed enthusiasm will ultimately be good, helping advance
the tech towards practical applications in trusted asset tracking, healthcare
data management, digital identity and smart contracts. But until more
applications are decoupled from requiring users to participate in the token
economy, widespread adoption will not happen. There simply isn’t enough trust
established yet.

A different kind of energy crisis

HPYE 2
/ SIGNIFICANCE 4

Marc Andreesen, co-founder of Andreesen
Horowitz famously said that “software is eating the world” in 2011 as he
discussed how software innovation was changing every industry.  The same could be said today about AI, but
with one letter added. Software now is Heating
the world as global scale data centers consume massive amounts of energy, with
nearly half of consumed energy used to keep them from overheating. The current
path is not sustainable for many communities that have fossil fuel power plants
at near peak capacity. Data centers are driving some of the larger new solar
deployments, particularly in the southwest. On the heels of energy dependence,
data centers are also going to consume a ton of water, as water-cooled systems
are on the rise (Google used 3.3B gallons of water to cool their data centers
way back in 2021, or about the annual need of a city of 100,000 residents).
Expect conflict in the future as public officials grapple with how to allocate
energy and water to big tech versus for our more mundane human consumption.

Disinformation security

HYPE 1
/ SIGNIFICANCE 5

When we first began connecting to the world
wide web, bad actors used viruses to try and steal data and corrupt systems.
From this, the anti-virus industry was born. As email became ubiquitous, it led
to a similar paradigm of spammers combated by anti-spam solutions. In the world
of AI, we have a crisis of misinformation and disinformation. Misinformation
used to train AI results in the proverbial garbage-in / garbage-out scenario
with sub-optimal AI results. But deliberate disinformation is creating even more
havoc, altering elections, changing perception about vaccines and public health
and worse. Expect 2025 to be the year when companies begin to offer the first
nascent “disinformation blockers”.  [I’d
call them anti-disinformation tools, but that sounds too close to a
double-negative]. These will work effectively, based on solid underlying
technology, but face an uphill battle to secure public trust.

Polyfunctional robotics

HYPE 1
/ SIGNIFICANCE 3

FIRST Robotics was founded by Dean Kamen all
the way back in 1989, but it has grown since to a global community of more than
615,000 annual participants across more than 100 countries. Conservatively,
1.5-2M past participants in FIRST high school robotics competitions are now out
in the workforce, and >85% of FIRST participants have pursued careers in
STEM. All of that excitement and expertise will reach a tipping point sometime
soon and it seems we are well-positioned for a paradigm shift from robots that are
purpose-built for one function, to robots that can do a variety of tasks. The
convergence of talent in industry, Agentic AI and autonomous systems like
Tesla’s autodrive make me think 2025 could be the year we begin to see truly
polyfunctional robots hit the main stage.

I’ll offer one more that I’m excited about,
but which is not yet ready for prime time – i.e. not a 2025 prediction.  Keep an eye out for developments, because you
just never know when AI might accelerate something more quickly than we
thought.

Quantum sensor networks

HYPE 0
/ SIGNIFICANCE – big, but probably not in 2025

Quantum mechanics can be used for more than
just computing.  It can also be used for
extremely precise sensing. Quantum sensors could revolutionize healthcare
imaging, with far more accurate and detailed images than we can get with
positron emission tomography (PET) or magnetic resonance imaging (MRI).  It could detect miniscule changes in earth’s
gravitational field suitable for predicting earthquakes and volcanic eruptions
before they occur and with sufficient lead time to get people to safety. They
can replace GPS for navigational positioning for self-driving vehicles and for
applications that can’t access orbiting satellites. Countries are pouring
billions of dollars into quantum research and quantum sensing has a chance to
deploy for some applications, perhaps sooner than full-on quantum computing.

I’d love to hear your feedback and
predictions.  Feel free to reach me at [email protected],
or join February 5th for RIoT’s annual State of the Region Address at Raleigh
Founded. The registration link is here.