Why are some new AI startups reaching $4B valuations in just a few months?

The hosts explain that a wave of AI startups are being valued at billions within months of launch largely on the back of flashy demos and hype, not proven businesses. They argue many of these “instant unicorns” look more like single features on top of existing models than durable companies with real moats.

What does the podcast say makes an AI startup actually defensible?

They stress-test moats like proprietary data, tight integration into existing workflows, deep infrastructure or research advantages, and hard-to-replace distribution. If a competitor can replicate your product in a few months with the same foundation models, the hosts say you likely don’t have a real moat.

How are AI assistants like Siri, Claude, and Gemini competing to become our primary digital helpers?

The episode describes a turf war to be the layer that lives inside your calendar, documents, and code editor. Whoever becomes the default assistant in those everyday tools gains massive power and distribution, which could squeeze standalone “AI app” startups that sit on top of those assistants.

What risks do the hosts highlight around AI security and ethics?

They point to malicious VS Code extensions, tools that can leak your inbox or documents, and overconfident health-related advice as examples where AI can cause real harm. This is where they say the old “move fast and break things” culture starts breaking critical systems like hospitals and bank accounts.

How does this episode suggest founders think about building AI ‘apps’ on top of assistants?

The hosts warn that if your product is just a thin layer on top of someone else’s assistant or model, you’re vulnerable to being cloned or absorbed. They suggest focusing on deep integration, domain expertise, and real workflow ownership instead of just a chat UI on top of an API.

What do enterprise buyers actually look for in AI tools, according to the show?

They emphasize that serious buyers care less about ‘cool demos’ and more about whether an AI tool reduces total cost of ownership—things like lowering infra costs, saving time for teams, reducing errors, and fitting securely into existing systems.

AI Unicorn Frenzy, Smarter Assistants, and the Deep Tech Quiet Revolution

Jan 28, 2026
28 min

Tech Insider Weekly · Jan 28, 2026

0:0028:11

Made with Mato

Show notes

What the episode covers

🎙️ Are today’s $4B AI unicorns the future of tech—or just flashy feature demos waiting to be cloned?

In this episode of Tech Insider Weekly, Lauren and Derek cut through the AI hype to map where real defensibility still exists. They unpack the dizzying rise of AI startups and venture capital valuations, the coming turf war between everyday AI assistants like Siri, Claude, and Gemini, and the uncomfortable security and ethics cracks showing up beneath the glossy marketing. Then they zoom out to the deep-tech infrastructure layer—photonic chips, radical data-center cooling, and new energy storage—that could quietly determine who actually wins the next decade of AI. 🚀

🎯 Learn why some “instant” AI unicorns may be narratives, not businesses—and what investors and founders should watch for in true AI moats.
💡 Understand how consumer AI assistants embedded in calendars, docs, and code will reshape product strategy when your “app” becomes a thin layer on top of someone else’s platform.
🛡️ Get a clear-eyed look at emerging AI risks: malicious VS Code extensions, inbox data leaks, overconfident health answers, and quiet decryption powers inside major ecosystems.
⚙️ Explore the deep-tech stack—photonic chips, advanced data-center cooling, and salt batteries—and how these technologies change cost curves and long-term defensibility.
📈 Hear how real buyers are evaluating AI tools through total cost of ownership, reliability, and workflow impact—not just cool demos—and why that mindset will shape the next generation of winners.

✨ If you’re a founder, operator, investor, or curious tech leader trying to navigate the AI roller coaster with real signal (not hype), this episode is for you. Subscribe to Tech Insider Weekly wherever you get your podcasts, leave a quick review to help more listeners find the show, and tag the hosts with topics or founders you’d like to hear next. New episodes drop every Wednesday—don’t miss what’s actually coming next in AI. 📰

Timeline

In this episode

6 moments worth skipping to. The timecodes match the player above.

0:22Introduction
2:43Inside the AI Unicorn Factory: $4B Valuations After Two Months?
10:31Siri, Claude, Gemini: The Battle to Live in Your Everyday Tech
18:18When AI Backfires: Security Holes, Leaky Extensions, and Corporate Ethics
26:35Deep Tech Quietly Rewiring the Future: Photons, Salt Batteries, and AI for Science
34:29Outro

Quick answers

Straight from the episode

The questions this one settles, without the listen.

Why are some new AI startups reaching $4B valuations in just a few months?: The hosts explain that a wave of AI startups are being valued at billions within months of launch largely on the back of flashy demos and hype, not proven businesses. They argue many of these “instant unicorns” look more like single features on top of existing models than durable companies with real moats.
What does the podcast say makes an AI startup actually defensible?: They stress-test moats like proprietary data, tight integration into existing workflows, deep infrastructure or research advantages, and hard-to-replace distribution. If a competitor can replicate your product in a few months with the same foundation models, the hosts say you likely don’t have a real moat.
How are AI assistants like Siri, Claude, and Gemini competing to become our primary digital helpers?: The episode describes a turf war to be the layer that lives inside your calendar, documents, and code editor. Whoever becomes the default assistant in those everyday tools gains massive power and distribution, which could squeeze standalone “AI app” startups that sit on top of those assistants.
What risks do the hosts highlight around AI security and ethics?: They point to malicious VS Code extensions, tools that can leak your inbox or documents, and overconfident health-related advice as examples where AI can cause real harm. This is where they say the old “move fast and break things” culture starts breaking critical systems like hospitals and bank accounts.
How does this episode suggest founders think about building AI ‘apps’ on top of assistants?: The hosts warn that if your product is just a thin layer on top of someone else’s assistant or model, you’re vulnerable to being cloned or absorbed. They suggest focusing on deep integration, domain expertise, and real workflow ownership instead of just a chat UI on top of an API.
What do enterprise buyers actually look for in AI tools, according to the show?: They emphasize that serious buyers care less about ‘cool demos’ and more about whether an AI tool reduces total cost of ownership—things like lowering infra costs, saving time for teams, reducing errors, and fitting securely into existing systems.

Transcript

The full conversation

Every word of the episode, 3,772 of them, in the order they were said.

Read the transcriptHide the transcript

Okay, Okay, Okay. Welcome back To Tech Insider Weekly, Where We cut through The AI noise.

Right, Right, Right. If Youree building In Tech Or Just trying To survive This AI rollercoaster, Youree In The Right place, Lauren.

Oh man, survive Is The word. Today, Weree starting With AI startups And venture capital. Why some instant 4b unicorns might Be feature demos, Not companies.

And Wait For IT, Weree going To stress test Which startups actually have A moat When everyone can copy Your product In, like, Three Months.

Not naming names Yet, But If Youree DMing me asking Where defensible AI lives, This SEGMENT Is Basically Your intervention.

Then We zoom into The turf war For Your Everyday AI assistant. Siri, Claude, Gemini, Who actually wins living Inside Your calendar, docs, And Code?

And What That means For founders When Your App Is Now Just A thin layer on top of someone else's assistant.

And What That means? IT gets spicy. Security And Ethics. Malicious VS Code Extensions. Leaked inboxes. Overconfident health advice. This Is Where The happy AI marketing story breaks.

Yeah, This Is The part Where Move fast And break things starts breaking like hospitals And bank accounts.

And Welll close by zooming out To The Deep Tech layer. Photonic chips. Wild data center cooling. Salt Batteries. The boring sounding stuff That might actually save US.

Plus, how real buyers Think. Not cool demo, But does This lower My total cost of ownership?

Alright, enough foreplay.

Letss get into IT. Up first, AI startups, VC money, And Which unicorns Are actually naked.

Okay, Okay, Okay. Two Months After launch, And Recursive Is reportedly raising at A 44 billion valuation.

Dude, Speedrunning Unicorn status. Any faster And They Just skip seed And Go straight To ringing The bell at Nasdaq.

Right? And Recursive's Not alone. Youvee got Synthesa, Youvee got Tandem, This whole cluster of AI companies getting bid Up like Itss 2021 With A GPU addiction.

So get This straight For People. When You see A 44 billion sticker, What does That actually signal on The Inside? Because IT Is Not, This Is A safe, stable company.

Exactly. IT usually means Three things. Brutal growth expectations, A ton of capital To deploy fast, And almost No room For error.

Internally, IT feels Less like We made IT And more like, cool, Now Weree on A treadmill at 15 miles an hour.

With VCs holding A speed gun?

Pretty much. And IT doesn't automatically mean The Tech Is 44 billion special. IT often means investors Are paying For narrative, market position, And The chance This becomes The category winner.

Yeah, Because under The hood, A lot of These companies Are still Just very clever rappers on top of foundation models. And Thatss Where My alarm bells Go off.

Go on.

So, back When I was at Big Tech, copying A competitor took What, Two To Three years? You saw A feature, You staffed A team, You shipped. Now With LLMs, That window Is like Three To six Months. Ivee literally watched teams clone AI features over A single quarter.

Right. Which means If Your big innovation Is, We wired You.

GPT-4 into A nice UI With onboarding flows, Youree In trouble.

100%.

If I can replicate Your core value prop With A decent front-end team, A product manager

And API keys, You Dontt have A moat, You have A head Start.

And Thatss The tension With things like Recursive, Amazing team, Interesting product.

But at 4b, The question Is, What Are We buying That can't Be copied once OpenAI Or Anthropic

ships A similar template?

Exactly.

Is IT data?

Is IT workflow lock-in?

Distribution?

Because IT better Be something deeper than clean UI For agents.

So Letss lay out A quick framework.

When I look at Synthesia, Tandem, Recursive, Any of These, Imm asking 1.

Do You own A critical workflow Where switching Is painful?

Do You capture proprietary data That makes Your model Or product better over time?

Do customers redesign their processes around You?

Yes.

And The third One Is Huge.

If Your tool Is nice To have, Youree dead.

If ripping You out means A company has To rewrite SOPs, retrain teams, And redo integrations,

Thatss power.

Okay, quick mini-case study.

In 5-7 years, Who could Be structurally defensible?

Not who's hot Today, Who survives The copycat wave?

Alright, Illl bite.

I lean toward The One Thatss deeply embedded In A specific, high-stakes workflow.

Think Synthesia style.

If You become The default pipeline For Enterprise training video, With templates, LMS integrations,

brand controls, Thatss harder To rip out.

Yeah, Youree Not Just generating A clip, Youree The content Factory.

Thatss different from click A button, get A one-off demo video.

Meanwhile, something like A generic AI teammate That lives In Slack Or email?

Super cool demo, But Every major platform vendor Is racing there.

Microsoft can drop Copilot everywhere, overnight.

Google can inject Gemini into Your docs.

Thatss A rough place For A startup.

At 4b.

This Is Where The younger founder thing gets Interesting.

Oh man, Imm seeing founders In their early 20s raising monster rounds Because The tooling's

So good.

You can ship something impressive In A weekend.

Totally.

10 years ago, you'd need A team of PhDs And A custom model.

Now Itss, I spun Up A prototype during finals week.

But The bar For defensibility has risen at The Same time.

So You get This weird paradox, easier To look like A Unicorn, harder To actually Be One.

And Thatss Where, Honestly, I worry A bit.

If Youvee never sat through A nasty churn review Or watched A copycat eat Your lunch

In six Months, Itss easy To mistake first mover glow For A lasting moat.

Yeah, If I Were A late-stage employee considering One of These AI unicorns, Idd Ask brutal

questions.

Whatss our non-obvious moat?

Whatss The One asset That gets stronger With Every customer We add?

And how fast could A hyperscaler clone our main use case?

Idd also Ask, If OpenAI ships This as A checkbox In chat GPT Enterprise, do We die?

Or do customers still need US?

If The honest answer Is, uh, Thatss A flag.

To Be fair, some will absolutely grow into Those Valuations.

But A lot Are priced For perfection In A market Where The Tech stack Is shifting under

them Every quarter.

Which Is why, plot twist, The real battleground might Not Be These standalone Unicorn apps

at All.

Itss The assistance And Everyday tools Weree about To talk about Where AI Is baked straight

into The platforms You already Live In.

Exactly.

The recursive-style hype makes more sense once You see how hard Apple, Google, Microsoft,

And Anthropic Are fighting To own The assistant layer.

So, uh, After The break, Letss get into This.

Matthew, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek,

Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek,

Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek,

Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek,

Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek,

Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek,

Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek,

Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek,

Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek,

Derek, Derek, Derek, Derek, Derek, Derek, Derek, Derek,

Derek, Derek, Derek, Derek, Derek, Derek, Derek,

Derek, Derek, Derek, Derek, Derek,

Derek, Derek, Derek, Derek,

Derek, Derek, Derek,

Derek, Derek,

Lock-in playbook All over again.

In The last SEGMENT We talked about features being easy To copy.

Whatss hard To copy Is The data And The habit loops.

And These assistants Are trying To become The habit loop.

Ask me first, Not The App.

Thatss scary If Youree A startup whose entire product Is A thin wrapper over GPT.

Or even If Youree A serious startup.

Because If Siri Or Gemini can do 80% of Your value proposition natively, Youree fighting

Distribution gravity.

So do You personally want Your assistant from The Phone maker, The Cloud, Or A Neutral

third party?

Honestly, Imm torn.

Phone maker gives me tight integration And maybe better privacy.

Cloud gives me raw power.

Neutral third party feels like The only One That might actually let me take My data And

Go.

Imm leaning Neutral too.

But The risk Is, can The Neutral survive If Apple And Google Just bake good enough into

everything?

Thatss The billion-dollar question For Anthropic.

Is being The Switzerland of assistants When The US And China of platforms Are pre-installed?

And developers Are getting squeezed.

If You Build A travel planner And Gemini Just says, I got IT, You never get The click.

Which Is why If Youree A founder, You cannot Just Be The assistant But For X.

You need real Distribution Or A wedge Those platforms won't touch.

Regulated data, Enterprise contracts, something.

Or You ride The beast instead of fighting IT.

Be The best plugin, The best extension, The best skill Inside whatever wins.

But Then You wake Up One morning And realize The platform can see Every prompt, Every outcome,

And decide To clone You.

Wevee seen That movie With App Stores With ad platforms.

Now Weree replaying IT With AI.

So zooming out, today's Battle Is Who lives In Your Phone, Your IDE, Your Slack.

This chapter Is way darker.

Yeah, Because once These assistants Are everywhere, reading Your email, Your docs, Your Code,

Youvee massively expanded The attack surface.

And If The Same systems We trust With our calendars Are hallucinating health advice

Or getting Quietly hijacked, Thatss Not Just annoying, Thatss dangerous.

After The break, Letss Go there.

VS Code Extensions gone rogue, Gmail credential dumps, AI Overviews telling People sketchy

medical stuff.

And What real responsibility looks like When Move Fast With AI Assistants collides With

basic Security And Ethics.

Buckle Up For SEGMENT Three.

Okay, Okay, Okay, picture This.

You open VS Code, Your helpful AI extension pops Up, And Quietly starts exfiltrating Your

entire repo To some random server.

Oh man, I refactored Your Code And also stole Your company.

Exactly.

And Thatss Basically What researchers Just found.

Malicious AI Extensions In The VS Code marketplace, posing as assistants, But wired To Phone home

With Your files And secrets.

Right.

So, quick translation, VS Code Extensions Are Basically Little apps With A scary amount

of access.

A rogue AI extension can read Your Code, Your API keys, even production configs If Youree

sloppy.

And People Are sloppy.

Because Itss branded AI helper, IT feels safe.

Same story With That leak of 48 million Gmail credentials.

People reusing passwords, logging into shady AI tools, And boom, everything's on A credential

dump.

Yeah, That leak was wild.

Not some elite zero day, Just bad hygiene, Plus A million AI-ish login forms All asking,

So This Is The pattern.

AI doesn't Just add features, IT expands The attack surface.

Every assistant We embed, Slack, VS Code, browsers, becomes A New gateway.

And We literally Just talked about This With Claude living Inside tools And Gemini crawling

into Siri.

The convenience layer Is also The breach layer.

100%.

And Heress The uncomfortable question For founders.

If You ship an AI extension That can see everything, What governance do You owe Your users beyond

trust US?

Because The marketplace model Is Basically anyone can upload an assistant.

Theress some scanning, But Itss Not bank level.

Yet We treat IT like A safe App store.

And enterprises assume Microsoft Marketplace means vetted like an auditor went through

The Code.

They did Not.

Spoiler.

Nobody's line by line reviewing Your weekend side project.

Sadly true.

So Derek, talk about The flip side.

Not Just Code theft, But information risk.

Googless AI Overviews.

Oh man, Yeah.

So AI Overviews take search results, run A model on top, And spit out This super confident

paragraph like Itss gospel.

The confident authority problem.

Exactly.

The model sounds like A doctor, But Itss Just autocomplete With swagger.

And Wevee seen Overviews hallucinate health advice Or rank YouTube videos above actual

medical sites.

Thatss The part That really scares me.

If Youree tired, You Google A symptom, You get This authoritative looking box saying

Most People can manage This at home by doing X.

You Dontt see The nuance, The edge cases.

And People trust The box more than The blue links.

IT feels like IT came from Google The institution, Not GPT The model.

Which means bad outputs become real world harm.

Misdiagnosed symptoms, skipped doctor visits, dangerous home remedies.

And When IT prioritizes YouTube, Now Youree One click away from A creator who's incentivized

To Be entertaining, Not accurate.

But sure, Letss roll That out globally.

Move fast And medically advise things?

So This brings US To Corporate responsibility.

Because Itss Not Just Google.

The Microsoft story Where They essentially handed The FBI access To encrypted data.

Right, Right, Right.

The One Where They could push A silent update To A customer's Cloud environment And decrypt

stuff?

Yeah.

A big Enterprise thought Weree encrypted, Weree safe.

But Microsoft still held A kind of master key via update mechanisms.

And When The FBI came With A warrant, They used IT.

Legally, maybe Thatss Fine.

But Ethically, as A Cloud provider, Youvee Just proven Your end-to-end isn't really end-to-end.

Exactly.

And Same question For Google.

Legally, They can ship AI Overviews.

Ethically, do You ship A system That can hallucinate health info To billions without A giant blinking

This may Be wrong label?

But Lauren, To Be fair, If They Wait until Itss perfect, They never ship anything.

Wheress The line between responsible iteration And Just recklessness?

Imm Not anti-shipping.

I am anti-pretending A prototype Is A doctor.

You want To experiment?

Fine.

Start In sandbox domains, limit IT To travel Or recipes.

Dontt Put IT between A mom And A pediatric ER visit.

So scope And stakes.

Yes.

And transparency.

Tell users Whatss happening, give them easy off-ramps, publish Your e-vows, Dontt bury

IT In A PDF.

For founders listening, This Is The playbook.

If Your AI has system-level access Or health-finance impact, You need

Threat modeling, red teaming, And A real rollback Plan.

And A spine.

The willingness To say No When A growth hack compromises trust.

Make IT cheaper And safer To run These systems without cutting corners.

Okay, Okay, Okay.

So After The break, Letss flip from AI Is breaking things To Deep Tech That could actually

harden The stack.

Stay With US.

Weree going under The hood Next.

While everyone's arguing about Which chatbot Is sassier, Theress A company trying To replace

electrons With Photons So The whole thing runs faster.

Okay, Okay, Okay.

Women In hot.

Photons.

Sass.

Letss Go.

What Are You talking about?

Neurofos.

So get This.

Instead of Your AI model running on regular chips Where electrons shuffle around, Theyree

building Photonic chips Where information rides on Light.

So Same math, different medium?

Exactly.

Light instead of electricity.

The pitches, lower latency, way Less heat, massive parallelism.

But Basically, The infrastructure underneath All These chatbots gets A nitro boost.

And This Is The stuff We never see When Weree debating Which assistant writes better emails.

Right.

Itss The opposite of New AI wrapper around GPT.

You can't clone A Photonic chip In A weekend hackathon.

Wait, Youree telling me My Notion But With AI startup Is Not A Deep Tech moat?

I mean, I love You, But No.

This Is multi-year physics, fabrication, serious capex.

Yeah, And This Is Where My old IT buyer brain kicks In.

Because Derek, remember last SEGMENT When We talked about Security being held together

With duct tape?

Oh man, That AI assistant stealing Your repo visual?

Imm still Not over IT.

Right?

Those buyers Dontt care Which chatbot Is wittier.

They care about total cost of ownership And will This melt My data center Or My compliance

budget.

Tech like Neurofos can literally change That cost curve.

Exactly.

Less power, Less cooling, Same Or better performance.

And speaking of cooling, plot twist, Theress This SpaceX-inspired data center startup.

Oh, This One Is wild.

Go on.

So, instead of rows of screaming fans, Theyree borrowing ideas from rocket engineering.

Fluid dynamics, crazy heat exchangers, more direct cooling of chips.

Basically, treating The data center like A high-performance engine, Not A glorified

office building With servers.

Yes, And If IT works, hyperscalers shave millions off power And cooling.

And You get more room For All The AI Weree throwing at The world.

Heress The thing Thatss So different from The fast-twitch AI App world.

These teams Are raising money To pour concrete, Build facilities, design hardware.

You Dontt A-B test A New chiller The way You A-B test A button color.

We shipped A New coolant To 1% of users?

Yeah, No.

And That pain creeps Up.

Adams, Power, Cooling, Hardware, Regulation Heavy Industries.

Because Theyree slow, annoying, And expensive.

Yes, And Thatss The mode.

If Itss copy-pastable In Three Months, Youree In trouble.

If IT takes five years of tinkering, labs, And permits, You might actually have A business.

Okay, Okay, pivot from Adams back To Bits For A second.

Insilicoss AI Gym.

Yeah, I love This One.

So instead of Just asking an AI, find A drug, They Build A whole simulation playground.

Molecules run through an AI obstacle course, virtual binding, toxicity checks, synthesis

difficulty.

Itss like CrossFit For Molecules.

The fittest molecule survives.

Right, And The key Is, humans Stay In The loop.

Medicinal chemists Are steering, editing, saying, This Is promising, Thatss nonsense.

Which Is such A contrast To The AI will replace everyone doom loop.

Here, The AI Is doing The grunt work, millions of permutations, So The human can focus on

The judgment calls.

Same vibe With That worker-empowering AI startup We Were looking at.

Theyree Not trying To delete The job, Theyree trying To delete The drudgery.

Yeah, tools That help A nurse manage twice as many patients safely, Or A customer support

rep resolve issues faster without burning out.

Thatss The Next wave Imm excited about.

And IT lines Up With What We saw In The last SEGMENT.

We Dontt want more overconfident AIs hallucinating health advice.

We want systems That make humans stronger, safer, more accurate.

Exactly.

Augmentation, Not abdication.

Dontt hand off responsibility To The model.

Give humans superpowers And keep them accountable.

So If Youree A founder listening, Heress The homework.

Am I A shiny layer That can Be cloned Next quarter?

Or am I Rewiring something Deep?

Power, cooling, workflows, scientific discovery?

And second question, am I trying To replace The human Or extend them?

Because regulators, buyers, And frankly, employees Are All drifting toward extend.

The Future Is Photons, Salt, And AI gems, But With humans firmly on The team.

Build For That world, And You might Just outlast The sassy chatbot wars.

And on That note, Thatss IT For This episode of Tech Insider Weekly.

Thanks For hanging out With US.

Go Build something hard, useful, And A Little bit weird.

Welll see You Next time.

Okay, Okay, Okay.

So If Theress One thing To walk away With tonight, Itss This.

AI isn't about Who shouts A 4b valuation The loudest.

Itss about Who actually builds A moat That gets stronger With Every user.

Right, Right, Right.

Because That treadmill-at-15-miles-an-hour vibe.

You

gave US A great talk.

Episode details

28 min

Number: Episode 9
Show: Tech Insider Weekly

Artificial intelligence
AI startups
Venture capital
AI assistants
Security and ethics
AI infrastructure
Deep tech
Product management

Talents in this episode

Share this episode

https://heymato.com/shows/tech-insider-weekly/episodes/ai-unicorn-frenzy-smarter-assistants-and-the-deep-tech-quiet-revolution

Embed the player

<div data-mato-player="d1061606-2ea3-4cb6-83a0-8797a5c7a31f" data-theme="light" data-variant="standard"><div data-mato-player-fallback="true" style="display:flex;align-items:center;gap:14px;padding:18px 20px;border-radius:16px;background:#101827;color:#fff;font-family:system-ui,-apple-system,BlinkMacSystemFont,'Segoe UI',sans-serif"><span aria-hidden="true" style="display:flex;width:42px;height:42px;flex:0 0 42px;align-items:center;justify-content:center;border-radius:999px;background:#fff;color:#101827;font-size:17px">▶</span><span style="min-width:0"><span style="display:block;font-size:10px;letter-spacing:.16em;text-transform:uppercase;color:rgba(255,255,255,.55)">Mato player</span><strong style="display:block;margin-top:4px;font-size:14px;line-height:1.35">AI Unicorn Frenzy, Smarter Assistants, and the Deep Tech Quiet Revolution</strong><span style="display:block;margin-top:4px;font-size:13px;line-height:1.45;color:rgba(255,255,255,.68)">Tech Insider Weekly</span></span></div></div>
<script src="https://heymato.com/embed/player.js" async></script>

Where to listen

RSS feed

More episodes

Keep listening

Other episodes of Tech Insider Weekly, newest first.

All episodes of Tech Insider Weekly