AI PULSE - Unpacking the Anthropic CEO Jack Clark Views on AI’s Future

TikTok Launches AI Alive to Animate Photos in Stories with Safety Features Audible is expanding its AI-narrated audiobook library Tesla's Optimus Robot Lags Behind Competitors Despite Ambitious Production Plans OpenAI wants to build a subscription OS for your life AI Strategies Spark Challenges and Criticism for Companies Like Klarna and Duolingo Unpacking the Anthropic CEO Jack Clark Views on AI's Future #AI, #Anthropic, #JackClark, #Tesla, #OpenAI, #AItechnology, #AInews

Episode Transcript

(00:00):
Welcome to Innovation Pulse, your quick, no-nonsense update on the latest in AI.

(00:09):
First, we will cover the latest news.
TikTok's AI Alive breathes life into photos.
Audible expands with AI narrations.
Tesla's Optimus progresses.
Open AI is personalized AI, and AI reshapes workplaces.
After this, we'll dive deep into Yakov Lasker's conversation with Jack Clark from Anthropik,

(00:32):
exploring AI's nuanced impacts and future possibilities.
TikTok introduces AI Alive, a new feature that transforms static photos into dynamic videos within TikTok stories.
Designed for creators of all skill levels, AI Alive uses intelligent editing tools to animate photos,
adding movement and creative effects.

(00:55):
Users can select a photo through the story camera and apply AI Alive to create engaging clips.
This innovation allows creators to easily bring scenes to life,
like turning a sunset photo into a cinematic experience or animating a group selfie.
AI Alive stories can be viewed across TikTok feeds and profiles.

(01:18):
The feature includes safety measures with moderation checks to ensure compliance with TikTok policies.
AI Alive stories are labeled as AI generated for transparency and include metadata to identify them as such.
This tool empowers users to enhance storytelling and creativity on the platform.

(01:40):
Audible Amazon's audiobook service is partnering with publishers to convert print and e-books into AI-narrated audiobooks.
This move aims to expand its catalog as it competes with Apple and Spotify in the booming audiobook market.
Audible uses Amazon's AI to offer publishers over 100 AI-generated voices in multiple languages and dialects.

(02:06):
The company plans to introduce a beta AI translation service for text and speech,
supporting translations from English to several languages.
Publishers can also opt for professional linguists to ensure translation accuracy.
In 2023, Amazon launched a beta voice tool for self-published authors to create audiobooks with synthetic speech.

(02:30):
Audible has over 50,000 titles labeled, narrated by Virtual Voice,
but the rise of AI narration has sparked debate about its impact on audio quality.
Spotify is also expanding its library with AI narration.
Join us as we explore the challenges facing Tesla's Optimus.

(02:53):
Tesla's Optimus humanoid robot is making strides, yet it still trails behind competitors.
CEO Elon Musk views Optimus as a multi-trillion dollar opportunity, forecasting the production of millions annually.
While Tesla deploys some robots in its factories, Musk anticipates selling them to customers soon,

(03:14):
pricing between $25,000 and $30,000.
However, progress remains in early development stages.
Recent updates showcase Optimus's improved walking and dancing abilities,
yet competitors like Unitry have demonstrated more advanced capabilities.
Milan Kovac, leading the Optimus program, noted ongoing improvements with simulation training.

(03:41):
Despite these developments, Optimus requires human assistance for practical tasks
akin to Tesla's self-driving technology.
The humanoid robot industry is expanding, with companies like Figure, X1, Unitry and X-Bang advancing.
While Tesla is progressing, skepticism remains about its leadership in the field,

(04:05):
especially given the reliance on human assistance during demonstrations.
OpenAI's Chief Executive Officer, Sam Altman, envisions developing an AI subscription similar to an operating system.
During the AI Ascent 2025 event, he shared ideas about creating custom AI models

(04:26):
that could integrate all personal experiences from conversations to emails.
Altman acknowledged the company's lack of a master plan, but emphasized a flexible approach,
focusing on improving AI infrastructure and developing smarter models.
He mentioned the potential for significant wealth creation through these developments.

(04:49):
Although details remain sparse, Altman expressed confidence in adapting strategies as needed.
He admitted that OpenAI cannot yet achieve the ideal personalized AI model,
which he describes as a system with extensive context of one's life.
The company aims to continue enhancing AI capabilities without a fixed plan,

(05:10):
indicating a dynamic, iterative process in its future endeavors.
AI is reshaping workplaces, but not without hiccups.
Klarna and Duolingo have been at the forefront of this shift.
Klarna initially embraced AI by freezing, hiring and replacing jobs with technology,

(05:31):
but now plans to hire more people, valuing human interaction for better service quality.
CEO Sebastian Siemiatkowski emphasizes investing in human support as the future.
Duolingo's recent AI-first approach is facing public backlash,
especially on TikTok, with users fearing job losses and reduced quality.

(05:55):
The company clarifies AI compliments, not replaces its experts.
Despite social media pushback, Duolingo's stock is thriving
and the company is optimistic about AI's role.
However, public concern persists.
Many fear AI threatens job security and reduces control over tasks.

(06:17):
A World Economic Forum study suggests many employers plan to rely more on AI,
fueling worries about its impact on the workforce and education value.
And now, pivot our discussion towards the main AI topic.

(06:38):
Hello everyone, Jakob Lasker here from Innovation Pulse.
I recently had the pleasure of diving into a fascinating conversation
between economist Tyler Cowan and Jack Clark, co-founder of Anthropik.
I've listened to countless AI discussions, but this one genuinely shifted my perspective.
Let me share why.
What makes Clark such a unique voice in the AI landscape is his unusual background.

(07:04):
Unlike most Silicon Valley technologists, he comes from journalism and the humanities.
This blend gives him a refreshingly balanced view that cuts through both the doom-mongering
and the techno-utopianism that dominate AI discussions.
First, let's talk economic impact.
While tech evangelists keep throwing around wild numbers like 20-30% GDP growth from AI,

(07:29):
Clark offers a sobering 3-5% estimate.
Now that got my attention.
Not because he's pessimistic about AI's capabilities.
Quite the opposite.
He's been repeatedly surprised by AI's rapid progress.
His caution comes from witnessing firsthand how devilishly difficult it is
to translate digital breakthroughs into physical world transformation.

(07:53):
Think about self-driving cars.
Remember when we were told they'd be everywhere by 2020?
That didn't happen.
Or consider robotics.
Clark mentioned a recent experiment with robotic hands that only achieved 60% reliability.
Would you let a robot with a 40% failure rate hold your baby?

(08:14):
Neither would he.
This digital to the physical gap is something we rarely discuss honestly in AI conversations.
What struck me most was Clark's perspective on which parts of our economy might resist AI transformation the longest.
It's not necessarily what you'd expect.
He believes specialized artisanal trades, like high-end gardening, will remain human-centered the longest.

(08:39):
Not primarily because of technical limitations, but because humans value the aesthetic judgment and personal connection.
We hire certain gardeners, not just for their skill, but for their taste and reputation.
And here's something counter-intuitive.
Clark thinks government agencies might adopt AI faster than many expect.

(09:01):
While Cowan suggested departments like Education would be among the last to embrace AI,
Clark argued that political pressure for efficiency and growth could accelerate adoption.
That's not something you hear often.
I was absolutely fascinated by their discussion about AI companions for children,
what Cowan called AI teddy bears.

(09:25):
As a parent himself, Clark admitted he's actually eager for these technologies.
Imagine having a well-meaning companion that could entertain your toddler while you're cooking dinner.
But this raises profound questions.
What happens when kids prefer their AI friends to human ones?
How do we balance monitoring with privacy?

(09:46):
Where do we draw boundaries?
Clark offered a compelling parallel between raising children with AI and developing AI itself.
In both cases, constant monitoring creates perverse incentives.
The child or AI learns to behave in ways that satisfy the monitoring system,
rather than developing authentic behavior.

(10:09):
Finding that balance will be one of our great challenges.
The conversation about media economics hit close to home for me as a content creator.
Clark acknowledges we're facing a fundamental restructuring.
In the old model, specialized journalism, like database coverage,
was cross-subsidized by content that attracted eyeballs.

(10:33):
Celebrity news.
But AI disrupts this entire ecosystem.
He envisions two distinct markets emerging.
Human creators using AI tools while maintaining personal connection with audiences,
think sub-stack, and expansive AI-extended fictional universes.
But when pressed about what happens to real-time news gathering and investigative reporting,

(10:58):
Clark candidly admitted he doesn't know.
That uncertainty is both refreshing and concerning.
Silicon Valley might be entering what Clark calls the era of manager-nerds,
people skilled at orchestrating AI systems rather than just coding them.
Companies like Mid Journey already operate with tiny human teams supplemented by AI agents.

(11:23):
The power will shift to those who can effectively direct these digital workforces.
When the conversation turned philosophical, exploring whether current AI systems are conscious,
Clark's thoughtfulness impressed me.
While some prominent figures like Jeffrey Hinton claim today's AIs are already conscious,
Clark takes a more nuanced view.

(11:44):
He describes current AI systems as existing in an infinite now.
They respond sophisticatedly to stimuli but lack a persistent sense of self or memory outside their context window.
They're on a trajectory towards something like consciousness,
but their current state resembles a tongue without a brain, responsive but not self-aware.

(12:07):
This philosophical question connects to very practical concerns about international relations.
What happens when countries increasingly rely on AI systems developed elsewhere?
If Peru turns over education, social security and defense to American AI systems,
what remains of Peruvian sovereignty?

(12:28):
Clark suggests governments will need to ensure AI systems reflect their population's values,
much like they do with other imported technologies and media, but with far deeper implications.
Some of Clark's specific predictions were mind-bending.
He believes we'll communicate directly with dolphins through AI translation by 2030.

(12:51):
He expects his child might live 130-150 years as AI accelerates medical breakthroughs,
and he thinks 10-year-olds might have the hardest time with the AI revolution.
They're a computer literate but caught in educational systems struggling to adapt.
What ultimately impressed me most about Clark was revealed in a simple detail at the interview's end.

(13:15):
When asked what he's currently learning, he mentioned two things.
Developing better tests for AI theory of mind and learning to juggle during walks.
That someone at the forefront of AI development finds meaning in mastering a purely human skill
speaks volumes about balancing technological advancement with human experience.

(13:38):
In an age where discussions about AI tend toward either apocalyptic warnings or promises of techno-utopia,
Clark offers something precious.
A grounded perspective informed by both technical knowledge and humanistic values.
His central question, how AI will change our material reality,

(13:59):
rather than just our digital experiences, cuts to the heart of what we should all be thinking about.
As we navigate this extraordinary technological transition,
his insights remind us that the most challenging questions aren't computational but human.
And that's something worth remembering as we build this future together.

(14:20):
This is Jakov Lasker for Innovation Pulse.
Until next time, stay curious and keep questioning.
And that wraps up today's podcast where we explore Tiktok's new AI Alive feature,
Audible's AI-Narrated Audiobooks, Tesla's Optimus Robot,

(14:42):
Open AI's Personalized AI Plans, and the evolving impact of AI in workplaces,
along with Jack Clark's insights on AI's economic and societal implications.
Don't forget to like, subscribe, and share this episode with your friends and colleagues
so they can also stay updated on the latest news and gain powerful insights.

(15:06):
Stay tuned for more updates.