Latest / Tech Talks With Kinsoft / Innovations in AI, Quantum, and 6G Technology 2025

Transcript

Innovations in AI, Quantum, and 6G Technology 2025 · Tech Talks With Kinsoft

• 0:00
  Welcome to Tech Talks with Kinsoft. This is where
• 0:02
  we cut through all the noise and really give
• 0:04
  you a shortcut to being properly well informed
• 0:06
  about, well, the crazy speed of tech today. It
• 0:09
  moves fast. It really does. And today we're doing
• 0:12
  a deep dive into some really exhilarating breakthroughs
• 0:14
  looking ahead to 2025. We're hitting AI, quantum
• 0:18
  computing, and the next wave of connectivity,
• 0:21
  6G. Yeah. Three massive areas. Exactly. And this
• 0:25
  isn't just about skimming headlines. We want
• 0:28
  to understand the foundational shifts happening
• 0:30
  now and what they actually mean for you. OK,
• 0:33
  let's unpack this. Sounds good. And it's an exciting
• 0:36
  time. Definitely. What's really interesting is
• 0:38
  how the picture forms when you look at all the
• 0:40
  different sources. We've got insights from, say,
• 0:42
  Moody's ratings, ENAX, big company news from
• 0:45
  Microsoft, Google. The usual suspects. Right.
• 0:48
  Plus cutting edge research, scientific communities,
• 0:51
  even some chatter on popular tech platforms.
• 0:54
  And together they paint this, well, coherent
• 0:57
  picture, but sometimes a surprising one of what's
• 1:00
  just around the corner. So not just theory. Not
• 1:03
  at all. We're really setting the stage to see
• 1:04
  how these innovations are actually moving out
• 1:06
  of the lab and are poised to genuinely reshape,
• 1:09
  well, our daily lives in whole industries. And
• 1:13
  speaking of reshaping industries, the first thing
• 1:16
  that just jumps out is this AI revolution. Specifically,
• 1:20
  autonomous AI agents really preparing to take
• 1:23
  the lead. Moody's Ratings is predicting some...
• 1:28
  Pretty significant strides for 2025. They're
• 1:30
  expecting these agents to plan, execute, and
• 1:33
  adapt tasks, all with minimal human intervention.
• 1:37
  That sounds, well, almost sci -fi. How exactly
• 1:40
  are these autonomous agents different from the
• 1:42
  AI we use now? What's the real game changer?
• 1:44
  Yeah, that's a really crucial distinction. Think
• 1:46
  of it less like just another smart tool and maybe
• 1:48
  more like an assistant that doesn't just understand
• 1:50
  your request but then figures out the steps needed.
• 1:52
  Actually executes those steps, maybe across different
• 1:55
  apps, and then learns from the result. All on
• 1:58
  its own. The game changer is that they could
• 2:00
  revolutionize industries by driving just unprecedented
• 2:03
  operational efficiencies and really turbocharge
• 2:07
  AI adoption everywhere. It's more than automation.
• 2:10
  It's about AI taking initiative. Okay, taking
• 2:13
  initiative. And the competition among developers
• 2:16
  is, well, it's intense, but it's not just about
• 2:19
  building the biggest model anymore. It's really
• 2:21
  shifted towards innovation and features, usability.
• 2:25
  Making it actually useful day to day. Exactly.
• 2:28
  The performance levels of the top foundation
• 2:30
  models, they've actually sort of converged in
• 2:32
  benchmarks. They're all getting incredibly good
• 2:34
  at accuracy, handling diverse tasks, and open
• 2:37
  source models. They're adding even more pressure,
• 2:40
  pushing everyone to innovate faster. So it's
• 2:42
  not just raw power. It's about finesse, intelligence,
• 2:45
  how it fits into our workflow. You mentioned
• 2:47
  enhancing the inference process maybe quickly.
• 2:49
  What does that mean for us? And this talk of
• 2:52
  synthetic data. Absolutely. So the inference
• 2:54
  process, that's basically where an AI model takes
• 2:58
  what it's learned, its training, and applies
• 3:00
  it to new data to generate a response. Making
• 3:03
  that better means making it faster, more efficient,
• 3:05
  more responsive. Which means quicker results,
• 3:08
  lower cost for businesses? Precisely. Quicker
• 3:11
  results, lower running costs, and synthetic data.
• 3:14
  Well, that's artificially generated data that
• 3:16
  mimics real world data. It's used to train AI
• 3:19
  when real data is maybe scarce or sensitive.
• 3:22
  Like medical data. Exactly. It shows real promise
• 3:25
  in structured areas like cybersecurity, healthcare.
• 3:28
  You can train models without compromising real
• 3:31
  patient privacy, for instance. But, you know,
• 3:33
  it's not a silver bullet. It still has limitations
• 3:35
  with complex, unstructured stuff like understanding
• 3:38
  natural language, the nuances of how we talk.
• 3:41
  Still best learned from real conversations. OK,
• 3:43
  that makes sense. So powerful tool, but you still
• 3:46
  need that messy real world data for like true
• 3:49
  depth. And we are seeing some striking real world
• 3:51
  impacts already. I mean, Mark Zuckerberg planning
• 3:53
  for meta to replace mid -level engineers with
• 3:55
  AIs this year. Yeah, that was a big headline.
• 3:57
  Or Coca -Cola, reinventing their Christmas ad
• 4:00
  solely using AI. Apparently, none of the shots
• 4:04
  were actually filmed on location. They're calling
• 4:06
  it an efficient way of saving time and money.
• 4:09
  It's clearly not just theory anymore, is it?
• 4:11
  Not at all. It's a fundamental shift in how major
• 4:14
  companies are thinking. Amazon's CEO confirmed
• 4:17
  AI will replace parts of their workforce, boosting
• 4:19
  automation, efficiency, Microsoft. Klarna, similar
• 4:23
  story. Raises big questions about jobs, of course.
• 4:25
  It absolutely does. But from the corporate side,
• 4:28
  that drive for efficiency is undeniable. And
• 4:31
  Microsoft, you mentioned them, they showcase
• 4:33
  some incredibly comprehensive AI stuff at Build
• 4:36
  2025. Their whole strategy seems to be about
• 4:39
  weaving AI right into the fabric of tools people
• 4:42
  already use every day. Which sounds smart, right?
• 4:44
  You make it like an invisible power up, not some
• 4:46
  new program you have to learn from scratch. Right.
• 4:48
  What's an example? Okay, take the Copilot AI
• 4:50
  upgrade. It's got this newfound autonomy in Windows
• 4:53
  11, Microsoft 365. It handles tasks using semantic
• 4:56
  search, so it actually understands what you're
• 4:58
  looking for, not just keywords. That's a leap.
• 5:00
  It is. It feels like it's anticipating your needs,
• 5:03
  not just reacting. Then you've got their AI powered
• 5:05
  surface devices. They're designed to learn your
• 5:08
  habits, anticipate needs, respond faster. The
• 5:11
  integration feels natural, not forced. That's
• 5:13
  key for adoption. Right. And interestingly, Microsoft
• 5:15
  is actually building its own proprietary AI models
• 5:18
  now. Really? Not just using others. Well, yes,
• 5:21
  partnering too. But building their own gives
• 5:23
  them much better control over response times,
• 5:26
  accuracy, privacy, how data is handled, and customization
• 5:30
  for specific things within their huge ecosystem.
• 5:33
  So it's about control and optimization. What
• 5:35
  about developers? Are they seeing benefits? Oh,
• 5:37
  absolutely. GitHub Copilot got major updates.
• 5:40
  Some are calling its debugging telepathic. Claims
• 5:44
  of cutting debugging time by up to 40%. That's
• 5:46
  a real tangible productivity boost. Huge. Beyond
• 5:49
  just tools, Microsoft is showing off real -world
• 5:52
  applications making a difference. AI analyzing
• 5:55
  patient data for personalized cancer treatments,
• 5:57
  processing weather data for better forecasts,
• 6:00
  providing real -time translation for hearing
• 6:02
  -impaired users. Full impact. Yes. And they're
• 6:05
  not doing it alone, these strategic partnerships.
• 6:06
  OpenAI, NVIDIA, even surprisingly, Elon Musk's
• 6:10
  Grok 3 .5. It's like assembling the Avengers
• 6:13
  of AI. Each bringing unique strength. That's
• 6:17
  a great analogy, the AI vendors. And speaking
• 6:19
  of powerhouses, we have to talk Google Gemini
• 6:23
  2 .5 Pro. That's being held as a... Pretty significant
• 6:27
  leap for 2025. What's setting it apart? Gemini
• 6:30
  2 .5 Pro is definitely setting new benchmarks,
• 6:33
  particularly its memory capacity. It can handle
• 6:35
  a million tokens. A million. That sounds like
• 6:38
  a lot. It's massive. It means it can effortlessly
• 6:40
  process huge data sets like entire books or company
• 6:43
  reports and maintain incredible accuracy like
• 6:45
  99 .7%. Even at that scale, imagine feeding it
• 6:49
  mountains of information and having it grasp
• 6:51
  the full context. Okay, that's impressive. And
• 6:52
  its deep research function transforms that data
• 6:55
  into simple... summaries, charts, saves huge
• 6:57
  amounts of time for analysis. It even did well
• 6:59
  on that humanities last exam benchmark 18 .8
• 7:03
  % score, which is impressive for AI. And it's
• 7:05
  number one on the El Marina leaderboard. That
• 7:07
  memory capacity alone sounds like a game changer
• 7:10
  for anyone dealing with big info loads. What
• 7:12
  about like creative stuff or practical uses beyond
• 7:15
  data crunching? Yeah, that's where Gemini Canvas
• 7:17
  comes in. It's multimodal, lets you seamlessly
• 7:20
  mix text, pictures, sounds. Think of it like
• 7:23
  a dynamic whiteboard where AI helps. you brainstorm
• 7:25
  and execute across different media. Very cool
• 7:28
  for creative and professional work. And for coders,
• 7:31
  its ability to write clean code, fix errors.
• 7:33
  It's completing something like 68 .6 % of full
• 7:36
  code edits. The business impact is already showing
• 7:39
  too. Online stores using it. Seeing 47 % more
• 7:42
  visitors, 32 % better sales. content creators,
• 7:46
  89 % more featured snippets, 63 % reduction in
• 7:49
  bounce rates. Wow. So these aren't just lab results.
• 7:51
  This is measurable business improvement. Exactly.
• 7:53
  Direct user empowerment, real business results.
• 7:56
  So zooming out again, the broader AI landscape,
• 7:59
  it really feels like we're heading towards this
• 8:01
  AI divide for businesses in 2025. Embrace it
• 8:04
  and thrive or get left behind. But apparently
• 8:07
  over 90 % of companies still need guidance. That
• 8:10
  tells us AI is absolutely splitting the industry.
• 8:13
  Its importance for productivity cannot be overstated.
• 8:16
  Think JPMorgan Chase checking 12 ,000 contracts
• 8:19
  in seconds. Seconds. That would take humans ages.
• 8:23
  Unbelievable. But here's the crucial insight,
• 8:26
  I think. AI works best when it's teamed with
• 8:28
  human skills. Doctors using AI to spot cancer
• 8:31
  faster in scans. Analysts using AI to sift market
• 8:35
  data for better decisions. So augmentation, not
• 8:38
  just replacement. Right. It's not AI replacing
• 8:40
  humans entirely in these complex roles. It's
• 8:43
  augmenting them. And this leads to the idea of
• 8:45
  AI evolving to provide results as a service.
• 8:48
  R -A -A -S. Results as a service. That's interesting.
• 8:51
  What's the distinction there? Well, R -A -S is
• 8:53
  a pretty... significant shift. Instead of just
• 8:55
  giving you data or a tool, AI results as a service
• 8:58
  means the AI itself delivers the finished, actionable
• 9:01
  answer, maybe even with recommendations. Give
• 9:03
  me an example. Okay, so instead of just analyzing
• 9:05
  sales data, it might actively suggest increase
• 9:08
  marketing spend in region checks by 15 % for
• 9:11
  a projected 5 % sales boost. And here's the reasoning.
• 9:14
  It delivers the value directly to the business
• 9:17
  user, makes AI more of a strategic partner. Got
• 9:20
  it. A clear differentiator. OK, from the cutting
• 9:23
  edge of AI, let's shift gears. Something even
• 9:25
  more mind bending. Quantum physics breakthroughs.
• 9:30
  Ah, yes. Buckle up. Exactly. Here's where it
• 9:33
  gets really interesting. How science is reshaping
• 9:36
  not just computing, but maybe our whole understanding
• 9:38
  of reality. We're seeing some truly foundational
• 9:40
  stuff. Take China's Zuzang supercomputer. 76
• 9:44
  qubit, photon -based. It did, reportedly, 2 .6
• 9:47
  billion years worth of computation in just four
• 9:49
  minutes. That number just doesn't compute. Literally.
• 9:52
  How? It uses a specific quantum method, Gaussian
• 9:56
  boson sampling. Without getting too physics -y,
• 9:58
  it's incredibly good at solving certain complex
• 10:00
  math problems, like designing new molecules and
• 10:03
  materials way beyond traditional supercomputers.
• 10:05
  Then there's Microsoft's Majorana 1 chip. First
• 10:08
  quantum ship with this new topological core architecture.
• 10:11
  Topological core architecture. Yeah. Sounds intense.
• 10:15
  What does that mean for reliability? These things
• 10:17
  are famously unstable, right? Great question.
• 10:20
  That's the whole point. Yeah. Think of it like
• 10:22
  building quantum computers on a fundamentally
• 10:24
  more stable foundation. This architecture uses
• 10:27
  something called a topoconductor, stay with me,
• 10:30
  to produce more reliable, scalable qubits. the
• 10:33
  basic building block, by making them less prone
• 10:36
  to errors from noise from the environment. So
• 10:38
  it means we're getting closer to practical quantum
• 10:40
  computing much faster, solving meaningful industrial
• 10:43
  problems in years, maybe not decades. A huge
• 10:47
  shift. And Continuum and Microsoft back in 2024,
• 10:50
  they actually created highly dependable logical
• 10:53
  quibbets, basically using multiple physical quibbets
• 10:56
  to make one super reliable one. This dramatically
• 10:58
  lowers error rates, huge implications for drug
• 11:01
  discovery, sustainable. energy materials, complex
• 11:04
  financial modeling. So we're moving from shaky
• 11:06
  experiments to something potentially useful for
• 11:09
  real industries. And quantum teleportation. Is
• 11:12
  that really happening? Like, beam me up, Scotty.
• 11:14
  Well, not quite like that. One source definitely
• 11:17
  reminds us human teleportation is pure science
• 11:19
  fantasy. But yes, in a quantum sense, Oxford
• 11:23
  researchers successfully teleported quantum information
• 11:26
  specifically. Logical quantum gates between separate
• 11:30
  processors on a... quantum supercomputer. Ah,
• 11:33
  information, not matter. Exactly. It's about
• 11:36
  instantly moving quantum states, which is critical
• 11:39
  for building larger interconnected quantum networks.
• 11:42
  And even Google's advances here, they have massive
• 11:45
  implications for cybersecurity. Quantum computers
• 11:48
  could potentially break current encryption methods.
• 11:51
  Which means we need new quantum resistant encryption,
• 11:54
  a whole new arms race. Pretty much. It's reshaping
• 11:57
  cryptography standards as we speak. Incredible.
• 11:59
  And it goes beyond just computing, right? Quantum
• 12:02
  physics is revealing, well, completely new aspects
• 12:05
  of reality itself. Oh, absolutely. Mind -bending
• 12:07
  stuff. Physicists created the world's first time
• 12:10
  quasicrystal. Inside a diamond, no less. It's
• 12:13
  a new phase of matter structured in time, not
• 12:15
  space. Structured in time. I can barely wrap
• 12:17
  my head around that. It's wild. And they found
• 12:19
  a new particle, semidirac fermions, in a material
• 12:23
  called zearses. It's massless moving one way,
• 12:27
  but gains mass moving another way. challenges
• 12:30
  our basic understanding of mass and energy, could
• 12:32
  lead to revolutionary applications, maybe like
• 12:34
  graphene did. Wow. And yes, scientists successfully
• 12:37
  turned light into solid matter, an astonishing
• 12:40
  breakthrough, plus an experiment where they engineered
• 12:42
  a photon that exists in 37 dimensions, pushing
• 12:46
  the limits of quantum mechanics. 37 dimensions.
• 12:49
  Okay. What does this all mean for how we think
• 12:52
  about reality? Well, you hear explanations like
• 12:54
  Neil deGrasse Tyson saying quantum tunneling
• 12:56
  is like the universe teleporting particles past
• 12:59
  barriers and pretending like it never happened.
• 13:01
  It's weird. Quirky but memorable. Right. And
• 13:03
  it brings us back to quantum entanglement, Einstein's
• 13:05
  spooky action at a distance. Linked particles
• 13:08
  affecting each other instantly no matter how
• 13:10
  far apart. It challenges our whole notion of
• 13:13
  local reality. Still baffling after all these
• 13:15
  years. It is. And then Brian Talks suggests the
• 13:18
  universe itself might behave like a quantum computer.
• 13:20
  These are the frontiers. Concepts like negative
• 13:22
  time or the universe as a simulation. They're
• 13:25
  more speculative, sure, but they grow out of
• 13:27
  these same bizarre quantum findings. Makes you
• 13:30
  really ponder the fabric of existence. It really
• 13:33
  does. Okay, from the fabric of existence, let's
• 13:35
  shift gears slightly to connectivity. 6G. Right.
• 13:39
  And XG. While 5G is still rolling out and maturing,
• 13:42
  6G development is accelerating. A whole new era
• 13:46
  of digital evolution. What can we expect, realistically,
• 13:49
  by 2025? Even if it's still early days. Even
• 13:53
  in development, 6G promises truly hyperspeed
• 13:55
  connectivity. They're talking maybe 100 times
• 13:58
  faster than 5G. 100 times. Yeah. Think truly
• 14:00
  instantaneous communication, seamless ultra -HD
• 14:03
  streaming, transmitting massive scientific datasets
• 14:06
  without a hiccup. Like real -time holographic
• 14:08
  meetings with zero lag. Okay. Holograms. Now
• 14:11
  we're talking sci -fi again. It's getting there.
• 14:13
  But beyond just speed, 6G networks will be intelligent,
• 14:16
  built -in AI. So the network itself can make
• 14:20
  decisions in real time, optimize data flow, predict
• 14:23
  user behavior, making it more energy efficient,
• 14:26
  more reliable. So a smarter network, not just
• 14:29
  faster. And security, with all that data flying
• 14:32
  around. Security and privacy are designed to
• 14:34
  take a massive leap. We're expecting enhanced
• 14:36
  encryption, maybe even quantum communication
• 14:38
  features. Better identity protection, really
• 14:41
  vital for finance, healthcare, sensitive sectors.
• 14:45
  Good to hear. It'll also enable truly seamless
• 14:47
  XR extended reality. Ultra smooth, highly immersive,
• 14:52
  augmented, virtual, mixed reality for gaming,
• 14:54
  education, collaboration. Imagine feeling like
• 14:57
  you're really in the same room as someone across
• 14:59
  the globe. Physical presence becomes less of
• 15:01
  a barrier. That could completely change remote
• 15:03
  work and interaction. What about all the things
• 15:05
  the IoT devices? Right. 6G will be the backbone
• 15:07
  for massive IoT ecosystems, handling huge data
• 15:10
  volumes with near zero delay. Smart homes, autonomous
• 15:13
  vehicles, whole smart cities. It's the invisible
• 15:17
  infrastructure making everything smart, responsive,
• 15:19
  connected. So it underpins a lot of other future
• 15:21
  tech. Absolutely. And for remote work and education,
• 15:24
  think real -time holographic meetings. As I said,
• 15:28
  AI -driven collaboration tools, uninterrupted
• 15:30
  virtual classrooms, more dynamic, more productive
• 15:34
  remote environments. Now, timeline, initial trials,
• 15:38
  maybe late 2020s, commercial rollout, probably
• 15:41
  early 2030s. So it's on the horizon, but the
• 15:44
  groundwork is being laid now. Exactly. Foundational
• 15:46
  work is happening right now. Okay. So let's pull
• 15:48
  back to the big picture. The overall tech industry
• 15:51
  outlook for 2025. We're seeing some serious growth
• 15:54
  predictions, right? Global IT spending projected
• 15:56
  up 9 .3 % in 2025, reaching $5 .74 trillion.
• 16:03
  Staggering numbers. It's huge. And a lot of that
• 16:05
  growth driven by focused investments. Generative
• 16:09
  AI is key, especially the supply -side infrastructure,
• 16:11
  the powerful hardware needed to run these models.
• 16:14
  The picks and shovels of the AI gold rush. Kind
• 16:16
  of, yeah. Cybersecurity is also massive. 53 %
• 16:19
  of organizations call it a top priority. And
• 16:21
  cloud computing continues its expansion, of course.
• 16:23
  Makes sense. But interestingly, even with all
• 16:25
  this spending, 92 % of companies... also plan
• 16:29
  cost -saving measures in 2025, like getting rid
• 16:32
  of redundant infrastructure so it's strategic
• 16:34
  spending, not just throwing money around. A balance,
• 16:37
  investment, but also efficiency. What about global
• 16:40
  confidence? Is everyone equally bullish on tech
• 16:43
  investment? Not quite uniformly. It looks like
• 16:46
  a bit of a mixed picture. Businesses in the EU
• 16:48
  seem a bit more cautious, while the U .S. shows
• 16:50
  a stronger willingness to invest right now. Interesting
• 16:52
  regional difference. And the most important technologies
• 16:55
  for companies in 2025, AI is still... still number
• 16:58
  one, followed closely by renewable energy and
• 17:00
  robotics. Widely adopted stuff already includes
• 17:02
  5G, 54%, open source software, 51%, IT automation,
• 17:07
  50%. And anything specific for Australia, since
• 17:10
  we're Kinsoft. Ah, yes. A couple of relevant
• 17:12
  points. Amazon AWS is planning a huge investment,
• 17:15
  AU $20 billion in Australian data center expansion
• 17:18
  between 2025 and 2029. That's a major commitment.
• 17:22
  Definitely signals long -term competence here.
• 17:24
  And another positive note, Australian businesses
• 17:26
  have apparently improved their average cyber
• 17:28
  attack recovery time by 38%. That's a significant
• 17:32
  jump in resilience. Shows focus is paying off.
• 17:35
  That is good news. Okay, what a fascinating landscape,
• 17:38
  bridging global trends and local specifics. So
• 17:41
  bringing it back to you, our listener, what stands
• 17:43
  out to you from all this, all these incredible
• 17:45
  advancements? How do you see these shifts impacting
• 17:48
  your life, your work, your decisions? I mean,
• 17:51
  the speed technology is integrating into everything.
• 17:53
  Everything. It's just remarkable. Yeah. And if
• 17:55
  we connect it all back, the big picture really
• 17:57
  highlights that the future isn't just, you know,
• 17:59
  coming someday. It's actively being built right
• 18:02
  now with these foundational technologies, AI,
• 18:04
  quantum, next gen connectivity, understanding
• 18:08
  these shifts. It's not just about curiosity anymore.
• 18:10
  It's really about empowerment. Empowerment. Good
• 18:13
  word. Empowerment to adapt, to innovate, maybe
• 18:16
  even thrive in this increasingly tech driven
• 18:18
  world we're all living in. Absolutely. Well,
• 18:21
  that was an incredible deep dive. Cutting -edge
• 18:23
  AI, mind -bending quantum physics, the future
• 18:26
  with 6G. We really hope you feel more informed,
• 18:30
  maybe a little bit inspired, perhaps slightly
• 18:32
  bewildered by the possibilities ahead. A healthy
• 18:35
  mix is good. Exactly. And as we wrap up, here's
• 18:38
  a thought. If quantum physics hints that our
• 18:41
  consciousness might collapse possibilities into
• 18:44
  different timelines, how might these rapid advances
• 18:47
  in AI and 6G influence the timelines we choose
• 18:50
  to create? personally, professionally. Something
• 18:54
  to ponder. Deep thought to end on. Why not? And
• 18:56
  of course, if today's deep dive sparked any ideas
• 18:59
  for your own businesses, IT, or security needs,
• 19:01
  maybe how these emerging technologies could give
• 19:04
  you an edge, we definitely encourage you to visit
• 19:06
  www .kinsoft .com .au. Let's talk about how they
• 19:10
  can be leveraged for your advantage. Absolutely.
• 19:12
  There's always more to learn, and we look forward
• 19:14
  to exploring it with you next time.