Best of LinkedIn: ICT & Tech Insights CW 41/ 42

00:00:00: This episode is provided by Thomas Allgaier and Frennus, based on the most relevant LinkedIn posts about ICT and tech insights from CW-FortyOne to FortyTwo.

00:00:09: Frenness support ICT enterprises in the form of delivering precise ICT market and pricing intelligence that analyst subscriptions and existing databases cannot provide.

00:00:18: you can find more info in the description right.

00:00:22: so

00:00:22: well actually imagine getting this mandate Like on a random Tuesday, right?

00:00:29: They just order your IT department to physically unplug Your company's web servers.

00:00:33: Just take them completely offline

00:00:35: immediately offline.

00:00:36: and You know that didn't happen in some dystopian tech thriller.

00:00:40: That actually happened this week two organizations that were running vulnerable PHP systems.

00:00:44: Yeah

00:00:45: It's wild

00:00:46: Right.

00:00:46: so for everyone tuning into This deep dive today whether you're like a digital transformation lead a CISO or just the tech strategist that visceral panic.

00:00:56: That is kind of the new baseline.

00:00:58: Absolutely

00:00:59: We're just officially out of that era where emerging technologies like AI or you know quantum computing are treated as these theoretical someday projects,

00:01:08: right?

00:01:09: Right it's kind Of Like Just around The Bend like one of These Days.

00:01:13: yeah Yeah I mean we Are Firmly in The Era of Hard Operational Accountability.

00:01:18: Now Right The Timeline for When These Advanced Technologies Actually Hit Your Daily Infrastructure It's Well, it's just collapsed.

00:01:24: We're seeing it across the board honestly

00:01:26: everywhere

00:01:27: exactly.

00:01:27: you've got governments issuing these hard non-negotiable deadlines for cryptographic overhauls.

00:01:32: You've got quantum hardware labs finally delivering independent measurable benchmarks which

00:01:38: is huge

00:01:39: It is.

00:01:40: and then on the other side you have AI actively weaponizing The software supply chain at a speed that basically makes traditional vulnerability patching completely obsolete.

00:01:48: Yeah completely obsolete.

00:01:50: So our mission today is to kind of cut through all the noise, All the press releases and figure out what actually working.

00:01:56: How it works And most importantly What you need do about It right now.

00:02:01: Let's get into it!

00:02:02: Lets start with a hardware side of this equation Because for years The narrative around quantum computing was just This brute force numbers game

00:02:09: Not totally Just cramming more qubits on a chip.

00:02:11: Yeah Vendors were racing To add more and more physical qubits But they were incredibly noisy Like super error prone

00:02:18: Right.

00:02:20: Based on some really interesting insights from Dr.

00:02:23: York Storm, and they were also posed by Antonio Gomez and Dixon Eriomaha The industry is pivoting hard to error correction

00:02:31: which is so necessary.

00:02:32: exactly So Google just published this big result in science regarding their new willow chip And they demonstrated this thing called magic state cultivation

00:02:42: Magic State Cultivation yet right?

00:02:45: So for the listeners who aren't quantum physicists?

00:02:47: what does he actual mechanism there was that mean?

00:02:50: Okay, so to understand magic state cultivation you really have to understand why quantum computers make mistakes in the first place.

00:02:56: okay because unlike classical bits which You know are strictly a zero or one qubits operate in these incredibly fragile Quantum states

00:03:05: super fragile.

00:03:06: Yeah They can be ruined by like a slight change in temperature Or even just stray electromagnetic waves in the room.

00:03:13: Wow

00:03:14: So to run really complex algorithms, you need these specific highly complex quantum states which they call t-states or magic states.

00:03:22: But creating them is incredibly messy.

00:03:25: Messy how?

00:03:26: Well previously You'd basically have to dedicate up till like ninety five percent of a Quantum machine's processing power just to constantly distill and purify These Magic States.

00:03:36: wait ninety five Percent

00:03:37: ninety Five percent Just filtering out the errors so that the remaining five percent of the machine could actually do The math you wanted it to do.

00:03:45: Okay, So It's like running this massive state-of-the art manufacturing plant.

00:03:50: right but almost all your energy and Your raw materials are just consumed repairing your own assembly line

00:03:57: Exactly

00:03:58: And at least almost nothing To build the actual product.

00:04:00: You want a sell?

00:04:01: That's A perfect analogy yeah!

00:04:03: How exactly did Google fix that Massive overhead on the Willichip.

00:04:07: Well, they optimized the cultivation process to a point where they cut those errors by forty times.

00:04:13: Forty times?

00:04:13: Yeah Wow!

00:04:15: They reached a ninety-nine point nine percent logical fidelity for these T states

00:04:20: Which is practically perfect

00:04:22: Right...they basically figured out how to identify and reject the flawed state much earlier in this distillation process before having a chance corrupting wider system.

00:04:32: Well, that makes sense.

00:04:33: Yeah and it completely shifts the economics of scaling a quantum computer because you just aren't wasting The vast majority of your hardware on error checking anymore

00:04:41: Right.

00:04:41: but you know Google isn't the only one out there claiming a massive leap in fidelity right now.

00:04:45: Oh definitely not

00:04:46: yeah.

00:04:47: So Jennifer Stravely and Marvin Lee highlighted Quantinium's new Helio system And this thing is already being prepped for deployment in Singapore.

00:04:54: right I saw that.

00:04:55: and then numbers they put out are wild.

00:04:58: They claim a logical error rate of five times ten to the negative sixth.

00:05:02: Which is tiny?

00:05:03: Yeah, they hit a ninety-nine point nine nine nine five percent logical fidelity.

00:05:08: and they said that you use something called a distance seven color code

00:05:11: To do it at distant.

00:05:12: Seven Color Code.

00:05:13: yeah

00:05:14: So break down that distance seven metric force.

00:05:17: what does that actually mean?

00:05:18: mechanically sure.

00:05:19: so A color code Is basically a way of organizing qubits into a specific geometric lattice

00:05:24: like a grid.

00:05:25: yeah think Of Like a honeycomb pattern.

00:05:27: It's set up so the qubits can monitor each other for errors without ever directly measuring the delicate quantum data they're holding.

00:05:34: Uh,

00:05:35: because measuring it would destroy the state?

00:05:36: Exactly!

00:05:38: So...the distance just refers to how robust that safety net is.

00:05:42: Okay In a Distance Seven code it will take chain reaction of at least seven simultaneous very specific physical errors To successfully corrupt logical data.

00:05:51: Got it

00:05:52: And Quantinium reported that every time they stepped up the size of this safety net, their errors fell by roughly twelve times.

00:05:59: See I'm looking at these Quantinum numbers and i just...I kind not buying it

00:06:03: Really?

00:06:04: Why Not?

00:06:05: Well

00:06:06: if you dig into the literature Amit Hagar actually published his really sharp fact check challenging there entire PR narrative on this.

00:06:14: Oh interesting

00:06:15: Yeah he points out that this ninety nine point nine nine nine five percent fidelity It didn't come from running a complex live algorithm.

00:06:22: Wait, what did they do then?

00:06:24: They literally generated a qubit left it completely idle for four seconds repeated that twenty-nine thousand times and Then called to the massive success.

00:06:34: Oh wow just left at idle.

00:06:36: yeah

00:06:37: That is essentially like parking car in a really secure garage For few seconds verifying it hasn't been stolen And then declaring it The most reliable vehicle ever manufactured.

00:06:46: Yeah, that's a bit of stretch.

00:06:48: And you know, Hager's critique actually goes even deeper than just the idle state.

00:06:52: Oh yeah?

00:06:52: Yeah he pushes back heavily on Quentinium's claim that they are a hundred times better then Google.

00:06:57: Right

00:06:57: because google was doing live math.

00:06:59: Exactly.

00:07:00: He points out this fundamental difference in how two companies handle errors.

00:07:04: Google ran one million rounds and their system decoded and corrected the error alive while the quantum circuit is still running.

00:07:12: But Quentinium ran a fraction of the rounds, extracted the data and then decoded the errors afterwards using classical computer.

00:07:20: Afterwards?

00:07:21: Which feels like a massive asterisk to me.

00:07:23: It's totally different ballgame yeah.

00:07:25: So

00:07:25: if you're a tech leader tuning into this And your trying decide which vendor to partner with or invest in They are all allowed define success Using completely different operational boundaries.

00:07:38: How do make rational decision?

00:07:40: Its tough.

00:07:41: We really need like a neutral referee on the field here.

00:07:44: Well, good news is we are finally getting one.

00:07:46: Jung So Kim noted that Sandia National Labs just introduced this thing called quaps?

00:07:51: Quaps...

00:07:53: Yeah, it stands for Quantum Operations.

00:07:55: It's an independent cross-platform benchmark that is specifically designed to measure fault tolerant readiness.

00:08:01: Oh!

00:08:01: That exactly what we need

00:08:02: Right...it normalizes the grading scale across systems from Google, IBM, Quantinium all of them.

00:08:08: so buyers can actually compare apples to apples

00:08:10: Which will inject a very heavy dose reality into market

00:08:14: Definitely.

00:08:14: And we're already seeing some industry leaders start to temper expectations, like Amy Forsey covered the Mizzou Quantum Day recently.

00:08:20: Oh yeah!

00:08:21: The IBM talk?

00:08:22: Exactly.

00:08:23: IBM Fellow Oliver Dial gave this really sobering assessment.

00:08:27: He stated plainly that useful general-purpose quantum computing is just not here yet

00:08:31: Not even close

00:08:32: Right.

00:08:33: His rule of thumb was super simple If a classical computer can solve a problem it absolutely should stay on a classical Computer.

00:08:40: Yeah because the overhead for quantum is still too high

00:08:43: Too High.

00:08:44: Progress has to be strictly modular and it has been proven piece-by-piece.

00:08:47: And you know that modular approach is actually driving a massive pivot in the industry right now?

00:08:52: How so

00:08:53: well if a single monolithic quantum processor isn't going to Be fully fault tolerant anytime soon, The focus shifts two main workarounds.

00:09:01: okay What are they?

00:09:02: physical networking and classical AI simulation.

00:09:06: So on the networking side Dr.. Chris now and Mihir Bhaskar shared this critical milestone from Ion Q.

00:09:12: oh i saw This one.

00:09:13: Yeah, they successfully linked a trapped ion qubit to a solid state quantum memory node at a rate of over a thousand connections per second.

00:09:21: Trapped Ion To Solid State?

00:09:23: That is like translating two completely different alien languages on the fly!

00:09:27: It really is...

00:09:28: How do they actually pass information between the two?

00:09:31: They use photons.

00:09:32: They basically map the fragile quantum state of the trapped ion onto a particle light send that photon through standard fiber optic cable and then translate it back into the solid-state memory And achieving a thousand connections per second, that's more than four times the previous record.

00:09:47: That's incredible!

00:09:49: It proves that quantum computing is going to scale the exact same way traditional data centers did you know?

00:09:54: Right...you

00:09:54: don't build one impossibly huge server.

00:09:56: Exactly You network specialized smaller modules together

00:10:01: and doing that physically Is actually getting easier.

00:10:05: Oh for sure.

00:10:06: Yeah, Mel Flayment pointed out that Cunex room temperature quantum memory was just named a time best invention of twenty-twenty six.

00:10:13: Room Temperature right?

00:10:14: That's the key.

00:10:15: yes Because instead of using those massive power hungry cryogenic cooling systems that have to get down to like near absolute zero They use warm rubidium vapor to hold the quantum states which is brilliant.

00:10:27: it really does.

00:10:28: Taking that extreme refrigeration requirement out of the supply chain makes stringing these modules together across a standard server rack infinitely more practical.

00:10:36: Absolutely!

00:10:38: By the way listeners, if you are enjoying this deep dive into how DeepTech is fundamentally rewiring enterprise infrastructure make sure to hit subscribe so you can catch future additions

00:10:47: for sure.

00:10:47: So okay, we have the physical networking path.

00:10:49: But the other major workaround right now is entirely digital.

00:10:53: We are essentially using today's AI supercomputers to simulate tomorrow's quantum hardware

00:10:58: Which was just a crazy concept.

00:11:00: it is Ellica Kiyoseva and Christos for highlighted this project running on Jupiter which is Europe's exascale super computing system.

00:11:08: Researchers from NVIDIA and Forchung Zentrum Yulich used up to eight thousand one hundred ninety-two GPUs, to run a one hundred four qubit simulation.

00:11:18: And what they were simulating was that exact magic state cultivation process that Google is working on physically.

00:11:24: This scale of that it's just incredible.

00:11:26: so they are Simulating the quantum physics in a totally virtual environment.

00:11:31: Yeah And doing it that way requires ninety percent fewer qubits than traditional simulation methods.

00:11:36: Which means engineers can actually figure out if a microscopic gate design works before they go and spend tens of millions of dollars trying to physically manufacture the silicon?

00:11:45: Exactly, an AI is taking very active role in this design process too.

00:11:51: Oh, really?

00:11:52: Yeah.

00:11:52: Dr.

00:11:53: Volkan Errol and Pat Manley discussed this joint project between Harvard, Anthropic & Quera.

00:11:59: They deployed large language models specifically CLAWD to design quantum circuits.

00:12:04: Claude is designing quantum circuits now.

00:12:06: Yes

00:12:06: The AI was tasked with routing the qubits And optimizing the gate sequences.

00:12:11: In seven out of eight core tasks The AI beat the absolute best human-designed configurations in published scientific literature.

00:12:19: It beat the human experts

00:12:21: By a lot.

00:12:22: We are talking about reducing gate counts by up to seventeen times.

00:12:26: AI can simply see those multi-dimensional routing efficiencies way better than human engineers Can.

00:12:31: okay, let me stop you there because if frontier AI models can optimize quantum circuit engineering by a factor of seventeen Doesn't that mean our assumptions?

00:12:40: About the cyber security timeline or completely wrong?

00:12:43: yeah right unquestionably all

00:12:45: right.

00:12:45: If ai is accelerating quantum hardware development this aggressively Doesn't that mean the window for quantum computers to break standard RSA encryption is shrinking drastically?

00:12:53: It is.

00:12:54: I mean, look we do not have a cryptographically relevant quantum computer today That's capable of breaking global encryption right

00:13:01: now Today but

00:13:02: applying exascale AI to quantum engineering that pushes the threat timeline Heavily to the left

00:13:08: which forces cybersecurity leaders to act Right Now.

00:13:11: The government Is Not Floating Suggestions.

00:13:13: Anymore They are handing down hard procurement mandates.

00:13:17: Yep, Sam Raymond detailed these deadlines set by the National Security Agency and they are strict.

00:13:24: By twenty-twenty seven all new commercial national security systems must natively support post quantum resistant algorithms.

00:13:31: Twenty twenty seven that's right around the corner.

00:13:33: And by twenty thirty All legacy systems That cannot support them Must be completely phased out.

00:13:38: See, that turned crypto agility from a nice little buzzword into a pass-fail procurement test for basically any enterprise doing business anywhere near the defense or critical infrastructure supply chain.

00:13:49: Totally!

00:13:50: Think about like a smart industrial controller on a manufacturing floor right?

00:13:54: You buy it in twenty twenty seven.

00:13:56: you're expecting.

00:13:56: a fifteen year life cycle makes sense but if its cryptographic protocols are hard coded into the firmware and they can't be updated to post quantum standards over You are going to be forced to physically rip out and replace that entire line of hardware long before it breaks.

00:14:13: And the nature of The Quantum Threat is actually expanding beyond just decrypting old files, too.

00:14:18: Chelsea Larsen-Andries brought up this really critical distinction.

00:14:22: Okay what's that?

00:14:23: Well we've heard endlessly about Harvest Now Decrypt Later.

00:14:26: you know the tactic where state sponsored attackers steal heavily encrypted databases today store them in a data center wait a decade for a quantum computer to unlock them.

00:14:37: Right data hoarding, that's the real threat

00:14:39: it is.

00:14:40: but she points out this new concept trust now forge later might actually be more immediately catastrophic.

00:14:48: Trust Now Forge Later.

00:14:50: explain that.

00:14:52: so future Quantum Algorithms specifically Shor's algorithm will be incredibly effective at breaking the cryptography behind digital signatures.

00:15:01: oh wow

00:15:01: Yeah, meaning an attacker wouldn't even need your old data.

00:15:05: They could dynamically forge software updates or impersonate you trusted cloud administrators Or sign malicious code as if it came directly from Microsoft and Apple.

00:15:14: That completely dissolves the foundation of digital trust in real time

00:15:18: Completely.

00:15:19: So what is the defense mechanism here?

00:15:21: Well Sharon Goldberg points out that post-quantum encryption Encryption is actually already viable and running at scale.

00:15:29: Oh, really?

00:15:29: Yeah cloud flare tunnel for example Is providing end-to-end post quantum encryption right now.

00:15:34: that's great.

00:15:35: It uses these complex mathematical lattices That even quantum computers struggle to solve which basically buys organizations crucial time For those legacy systems.

00:15:43: they can't just upgrade immediately.

00:15:46: But you know as CISO scramble define the solutions They are being bombarded by vendor hype especially around this different technology called Quantum key distribution or QKD?

00:15:56: Oh,

00:15:56: yes.

00:15:57: Yeah qkd

00:15:58: yeah.

00:15:59: adi lamar offered this brilliant framework for cutting through all that noise.

00:16:03: she insists That you ask three killer questions during any qk d pitch.

00:16:07: i love these questions.

00:16:08: they're so good.

00:16:09: first Does the system actually authenticate The endpoints Or do You still have to rely on vulnerable classical cryptography just To prove who was On the other End of the line

00:16:18: right?

00:16:18: because if you Do what's the point

00:16:20: exactly?

00:16:20: second What published independent assurance standard, has the hardware actually been evaluated against?

00:16:27: Most of time.

00:16:28: The answer is none crickets

00:16:29: Yeah.

00:16:30: and the third question she raises Is the absolute deal breaker.

00:16:34: What is the maximum distance this signal can travel without requiring a trusted node?

00:16:39: This is the big one.

00:16:40: yeah

00:16:41: because QKD relies on sending delicate single photons over fiber optics.

00:16:46: But those photons degrade over distance

00:16:48: right.

00:16:49: To go farther than roughly a hundred kilometers, the signal has to enter a repeater node.

00:16:54: Be completely decrypted into plain text and then re-encrypted onto a new photon which

00:16:59: is just some massive fatal flaw.

00:17:01: Massive!

00:17:02: It's like chaining top secret briefcase your wrist right?

00:17:06: But every fifty miles you have to hand it to some random security guard in a toll booth who opens it reads the contents puts it on a new brief case passes along

00:17:14: exactly.

00:17:15: If you don't physically own and secure every single one of those repeater nodes, your physics-guaranteed secrecy is just a complete illusion.

00:17:24: It's exactly why the NSA, Britain's National Cyber Security Center, France's ANSSI ,and Germany's BSI have all issued formal positions stating they prefer math based post quantum cryptography over physics base QKD.

00:17:39: it Just makes more sense.

00:17:40: yeah

00:17:40: You can run PQ encryption Over the internet infrastructure we already have

00:17:45: Which brings up this massive geopolitical issue regarding who actually controls the submerging infrastructure.

00:17:51: Oh, yeah.

00:17:52: Bozanis Lominsky noted that twelve point six billion dollars was invested in quantum startups and twenty-twenty five.

00:17:58: That's

00:17:58: a lot of money

00:17:59: But sixty percent of that capital was concentrated on just top ten deals.

00:18:04: The consolidation there is staggering.

00:18:05: We are basically talking about laying the foundational digital roads and bridges for the next century of computing.

00:18:11: And only a handful of megacorporations, sovereign wealth funds will own the IP in cloud access points.

00:18:18: Exactly!

00:18:19: Speed of adoption without maintaining data sovereignty is just faster way to lose control your technological future.

00:18:26: You know we don't even have.

00:18:27: look ahead until the twenty thirties.

00:18:29: see what happens when this speed of technological exploitation breaks our security paradigms.

00:18:34: No

00:18:34: We really don't.

00:18:35: Because AI is doing it to our classical day-to-day infrastructure right now,

00:18:39: which circles us back into that reality check we started the show with.

00:18:43: Bob Carver makes this compelling argument that traditional vulnerability management is just

00:18:48: dead and gone.

00:18:49: You can no longer treat as a patching backlog where you prioritize critical updates over like.

00:18:56: AI has accelerated the research discovery and weaponization cycle so aggressively that The only metric that actually matters now is the time between a confirmed risk, and verified mitigation.

00:19:07: Exactly

00:19:07: CISOs have to shift from tracking patch counts To measuring their exploitable exposure in real-time

00:19:14: And Huzefa Ono and Aaron Okdora detailed exactly what that Accelerated weaponization looks like in practice.

00:19:21: security teams recently faced this brutal twenty four hour window perfectly demonstrates the new baseline.

00:19:28: In a single day, researchers disclosed two critical PHP vulnerabilities collectively known as PHP to Shell and they carried

00:19:40: Let's contextualize a nine point two out of ten for a second.

00:19:43: That means it allows for unauthenticated remote code execution.

00:19:47: Yes,

00:19:47: so an attacker does not need a password.

00:19:49: They do not need an account on your system.

00:19:51: they can just remotely send a command over the internet and take complete administrative control Of your server.

00:19:56: exactly And on that exact same day A light LLM supply chain compromise dropped containing a malicious info stealer alongside Ten different no.js vulnerabilities.

00:20:07: In the past, attackers might take weeks to build exploits for these flaws but now using AI assistance.

00:20:12: These vulnerabilities were being actively exploited in the wild within hours of disclosure.

00:20:18: Hours?

00:20:20: That is why Cynthia Kaiser pointed out that CESA – The Cyber Security and Infrastructure Security Agency actually resorted to ordering organizations to take their PHP systems entirely offline until patches could be applied.

00:20:33: Right

00:20:34: When the federal government tells you to physically unplug your infrastructure, The defensive window hasn't just shrunk it has completely collapsed.

00:20:42: It creates a massive imbalance.

00:20:44: AI gives the offense an overwhelming advantage and defenders are rightfully asking You know who is responsible for this ecosystem chaos?

00:20:51: Who was holding them accountable?

00:20:53: Right?

00:20:53: Brian Panarello sharply criticized openAI regarding This exact dynamic.

00:20:57: Oh

00:20:57: what did he say?

00:20:58: Well OpenAI recently tried To claim public relations credit for giving defenders time to prepare.

00:21:04: For emerging threats, but simultaneously they are the source platform attackers use to generate zero day vulnerabilities.

00:21:11: Wow!

00:21:12: Panarello argues that responsible disclosure needs to be rooted in actual vendor accountability.

00:21:17: not just.

00:21:18: you know PR framing.

00:21:19: I mean he's not wrong.

00:21:21: so how does a compliance or security department actually adapt to this?

00:21:25: because

00:21:29: No, you integrate it into the frameworks that are already used.

00:21:32: Santosh Kudzar highlighted a crucial advisory from Set-In that explicitly flags AI agents themselves as independent attack vectors.

00:21:40: That makes sense!

00:21:41: Yeah if an AI agent has autonomy to execute code or access databases... It must be locked down with strict least privilege access.

00:21:50: You just cannot trust the AI to govern its own permissions.

00:21:53: Never

00:21:53: And Maria Luisa Radondo Velazquez offered really practical solutions for compliance teams.

00:21:58: here She pointed to the new ISO IEC-II-C-O-NINET guidance.

00:22:03: Oh, The New ISO Standard?

00:22:05: Yeah

00:22:05: it is designed To take highly specific AI threats like data poisoning where attackers silently corrupt their training data or prompt injection Where they trick the AI into bypassing its own safety rules and It maps them directly Into existing ISO-IIi security frameworks.

00:22:23: That's incredibly helpful.

00:22:24: It is.

00:22:25: You don't have to invent a whole new security language, you just adapt the checklists your auditors already rely on Right

00:22:31: And we also have to fiercely protect the physical layer.

00:22:34: while were at it Robert M. Lee shared that Dregos Accenture and SANS have launched A ten million dollar foundation To offer free operational technology cyber-security resources Software AND training specifically to small U.S water & power utilities.

00:22:50: Which is so critical

00:22:52: Because when AI accelerates a digital attack against a local water treatment plant, the consequences aren't just lost data.

00:22:58: They are physical.

00:22:59: they're immediate and potentially lethal.

00:23:02: We've covered a massive amount of ground today And I want to leave you with a final thought on the paradox of control we find ourselves in right now.

00:23:09: Yeah it's

00:23:10: quite a paradox.

00:23:11: On one hand...we have the brilliance Exascale supercomputers and AI models to engineer the atomic layout of quantum processors.

00:23:19: Processors that will eventually shatter the mathematical foundations of global secrecy, right?

00:23:25: We are architecting in future with unbelievable precision yet on any given Tuesday we're still completely at the mercy.

00:23:33: a broken supply chain frantically unplugging web servers because can't deploy PHP patch faster than an AI script.

00:23:39: weaponize it

00:23:40: exactly.

00:23:41: The

00:23:42: true existential risk for tech professionals isn't just the sheer capability of quantum computing or generative

00:24:11: AI.