OpenAI-Scale AI Split, Apple's Generative AI Chips, and Amazon's Shopping Revolution

In this episode, Bob delves into OpenAI's decision to end its partnership with Scale AI, following Meta's investment, and the broader industry implications of this move. The discussion includes OpenAI and Scale AI's responses to these partnership changes. Bob then highlights Apple's integration of generative AI in chip design, discussing its significance for the industry. The episode shifts focus to the rise of agentic AI tools and the increasing market competition they foster. Bob also explores Amazon's innovative "buy for me" agent, which is set to revolutionize online shopping. Lastly, breakthroughs in AI brain activity simulation by Martin Schrimpf are examined.

Episode Transcript

(00:00):
OpenAI is cutting ties with Scale AI afterMeta's multi-billion dollar investment in the

(00:05):
startup.
Welcome to The AI News Daily Brief, your go-tofor the latest AI updates.
Today is Thursday, June 19, 2025.
Here’s what you need to know about the shiftingdynamics in AI data partnerships.
Let’s dive in.

(00:25):
Imagine having a reliable partner you've workedwith for years, and suddenly, they team up with
your biggest competitor.
That's the situation OpenAI finds itself inwith Scale AI after Meta's recent investment in
the data labeling company.
This move has sparked a wave of speculationabout the future of Scale AI and its
relationships with major AI players.

(00:49):
The importance of this story lies in thebroader implications for the AI industry.
OpenAI's decision to drop Scale AI hints at agrowing trend among AI companies to seek
"neutral" partners for their data needs.
With companies like Google reportedlyconsidering similar moves, it seems the race is
on to secure data providers that can offerunbiased support.

(01:13):
An OpenAI spokesperson told Bloomberg that thecompany was already winding down its work with
Scale AI, seeking more specialized dataproviders to develop increasingly advanced AI
models.
This shift reflects a strategic pivot as OpenAIlooks to maintain its edge in a rapidly
evolving landscape.
Adding to the intrigue, Scale AI's generalcounsel has publicly stated that Meta will not

(01:38):
receive preferential treatment, and thatconfidential information from other customers
will remain protected.
Despite these assurances, the perception ofbias has led some of Scale AI's biggest clients
to start exploring alternatives.
In response to the upheaval, Scale AI's interimCEO Jason Droege announced the company's

(01:59):
intention to "double down" on its applicationsbusiness, focusing on building custom AI
applications for governments and enterprises.
This strategic shift could redefine Scale AI'srole in the industry as it navigates these
turbulent waters.
Apple is making waves in the tech world byplanning to use generative artificial

(02:20):
intelligence to speed up the design of itscustom chips.
This comes directly from Johny Srouji, Apple'ssenior vice president of hardware technologies,
who shared these insights during a speech inBelgium last month.
He was there to receive an award from Imec, aleading semiconductor research and development
group.
Srouji's remarks highlight Apple's ongoingcommitment to leveraging the most advanced

(02:43):
tools available to enhance its chip designprocess.
Imagine being on the cutting edge oftechnology, always pushing for innovation.
That's exactly where Apple stands with itsapproach to chip design.
Srouji took the audience on a journey from thefirst custom A4 chip in an iPhone back in 2010
to the latest innovations powering Maccomputers and the Vision Pro headset.

(03:07):
His speech underscored how crucial it is forApple to use cutting-edge tools, like the
latest software from electronic designautomation firms, to stay ahead in the game.
Why does this matter?
Well, the two leading companies in theelectronic design automation industry, Cadence
Design Systems and Synopsys, are racing tointegrate artificial intelligence into their

(03:30):
offerings.
Srouji emphasized that generative AI techniquescould significantly boost productivity by
allowing more design work to be completed inless time.
It's a huge deal because it means Apple cankeep innovating faster and more efficiently.
In his speech, Srouji also shared an importantlesson from Apple's past: the importance of

(03:52):
making big bets.
He pointed to Apple's decision to transitionits Mac computers from Intel chips to its own
Apple Silicon in 2020 as a prime example.
This was a massive gamble with no backup plan,but it paid off, allowing Apple to fully
integrate its hardware and software.
The competition to develop the most advancedagentic artificial intelligence is heating up,

(04:15):
and companies like Meta Platforms and Amazonare emerging as early winners in this race,
according to analysts from Bank of AmericaSecurities.
This development is drawing significantattention because agentic AI tools are capable
of not just chatting but taking actions online,like making purchases or scheduling
appointments.

(04:37):
Imagine having a digital assistant that doesmore than just answer your questions—it
actively helps you manage your day-to-dayactivities.
That's the promise of agentic AI, and it's whytech giants are investing heavily in this area.
OpenAI launched its Operator earlier this year,and Google followed suit with its Mariner
agent.

(04:57):
It's a race to become the go-to "agent" thathelps consumers navigate their lives.
Why should you care?
Well, analysts believe that while these AIassistants are not yet ready to handle all of
our daily tasks, the potential for growth isenormous.
Companies like Meta and Amazon are poised tobenefit significantly as this technology

(05:20):
evolves.
Meta's AI assistant, already boasting over abillion users, is seen as a central component
of the company's future, integrating seamlesslywith its apps to offer a unified interface for
content, commerce, and productivity.
Justin Post, an analyst with Bank of AmericaSecurities, highlights Meta's AI assistant as a

(05:43):
potentially underappreciated asset within thecompany's ecosystem.
He suggests that as a general-purpose agent,Meta AI could become a primary tool for users,
driving content discovery and even automatingsome advertising campaigns.
This is especially timely as Meta is investingheavily in its AI initiatives, recently

(06:04):
committing over fourteen billion dollars toScale AI.
Meanwhile, Amazon is testing an intriguingfeature called a "buy for me" agent, which
could revolutionize how we shop online.
This agent might purchase items unavailable onAmazon.com on behalf of customers, streamlining
the shopping experience.

(06:24):
As agentic AI enhances price transparency andsimplifies product discovery, Amazon is
well-positioned to leverage its strengths inpricing, selection, and delivery, areas where
it already excels.
The stakes are high, and the competition isfierce.
As these AI assistants become moresophisticated, companies like Google could face

(06:47):
"disruption risk" in their core businesses,like search.
However, for Meta and Amazon, the potentialrewards seem promising, with analysts rating
their stocks as buys, projecting significantgrowth as they capitalize on these advancements
in agentic AI.
The world of artificial intelligence is on thebrink of a breakthrough that could change how

(07:09):
we understand the human brain.
Martin Schrimpf, a leading figure in thisfascinating field, is working to build AI
models that can not only mimic but alsoinfluence brain activity.
Welcome to The AI News Daily Brief, where weexplore how AI is reshaping our understanding
of the mind.
Imagine a future where we can precisely controlbrain activity using AI-generated sentences.

(07:36):
That’s what Schrimpf's groundbreaking researchhas achieved.
By developing AI models that simulate humanneural processes, Schrimpf and his team have
created a way to alter brain activitynoninvasively, marking a milestone in
neuroscience.
Why should this matter to you?
Because this technology opens doors topotentially treating conditions like depression

(07:58):
and dyslexia by tailoring stimuli to individualbrain patterns.
It’s not just about understanding thebrain—it’s about using that understanding to
improve lives.
Schrimpf’s approach is innovative.
He uses artificial neural networks, which areinspired by the way neurons in our brains
communicate, to create what he calls a "digitaltwin" of the brain.

(08:21):
This involves testing people on tasks relatedto language and vision, then refining AI models
based on their brain activity to make them morehumanlike.
The potential here is enormous.
Schrimpf’s open-source platform, Brain-Score,is already a treasure trove of neural and
behavioral data, helping researchers worldwidetest AI models against human data.

(08:46):
This collaborative effort is pushing theboundaries of what we know about the brain.
But with great power comes greatresponsibility.
Schrimpf acknowledges the ethical implicationsof his work, emphasizing the need for careful
consideration of how these models mightinfluence human thought.
It’s a delicate balance between innovation andethics, one that he is committed to navigating

(09:09):
responsibly.
That’s it for today’s AI News Daily Brief.
As AI models inch closer to unlocking thesecrets of the human brain, the possibilities
for medical and technological advancements areendless.
Thanks for tuning in—subscribe to stay updated.
This is Bob, signing off.
Until next time.