BIG TECH

Why Apple needs to fight AI brain drain

Meta is continuing its poaching spree. 

Ke Yang, who led Apple’s initiative to develop AI-powered search, is reportedly joining Meta, according to Bloomberg. Yang headed Apple’s  Answers, Knowledge and Information team, which aimed to give Siri capabilities akin to ChatGPT, aiming to help it compete in the growing AI search market. 

The move marks yet another in Meta’s effort to build up its stable of AI talent. The company has poached researchers from OpenAI, Google DeepMind and Thinking Machines Lab in recent months, courting talent with eye-popping compensation packages and promises of superintelligence development.

This, however, may be a blow to Apple in particular as it struggles to find its place in the AI race. Though the iPhone maker has vowed to spend $500 billion on AI over the next four years, it has so far little to show for it compared to competitors. Catching up could require a significant pivot, Mahmoud Ramin, research director at Info-Tech Research Group, told The Deep View. 

“It may now be the time for Apple to evolve into an “AI-native” company rather than remaining primarily a hardware and software integrator,” Ramin said. 

There are a few things Apple can do to course correct, Ramin said.

•  Boosting incentives to retain talent and creating a stronger recruitment pipeline could allow it to fight its “talent drain,” he noted. 

• Additionally, Apple should provide an “open research environment” to offer its teams more autonomy to explore AI, said Ramin. 

• And if all else fails, Apple can choose to buy over build, acquiring other AI-first companies and forging partnerships to build up its stable. (This may already be in the works, as Apple is reportedly nearing a deal to acquire engineers and technology from computer vision startup Prompt AI.) 

Apple continuing as it has been could come with serious consequences, said Ramin. “Falling behind in AI enablement provides an opportunity for competitors to embed more powerful AI capabilities throughout their operating systems and devices, making them a stronger choice for customers.”

Apple has long taken the “wait and see” approach with technology development, taking the ideas of other tech firms and refining them with the Apple hallmarks of a sleek user experience and a walled garden of privacy. But in taking that strategy with AI, Apple may be shooting itself in the foot. By the time it develops its own AI – no matter how sleek, refined or Apple-ified it is – competitors will have already developed far stronger models. 

It’s clear that AI development is a cross-industry effort, with companies like OpenAI, Nvidia, Oracle and Meta forging billion-dollar partnerships at a rapid clip. Apple may need to take a page out of its playbook to finally catch up.

TOGETHER WITH IBM

Data in the driver’s seat

With IBM, Scuderia Ferrari HP focused on data to fuel strategy, precision, and passion. On the track and beyond, data became an engine of innovation for fans.

Ferrari’s fan app transformed raw race data into immersive, real-time storytelling, including:

• AI-generated race summaries from live and historical data

• Scalable hybrid cloud infrastructure for enhanced speed and agility

• Real-time personalization based on fan behavior

Every lap, corner and millisecond becomes a window into the data-driven decisions that brings fans closer to the team like never before.

Know more

RESEARCH

Google takes AI to the lab

While some tech firms are busy making AI automate tasks, handle shopping and act more … romantic, Google is taking it to the lab. 

On Wednesday, Google released Cell2Sentence-Scale 27B, or C2S-Scale, a 27 billion parameter model to “understand the language of individual cells.” The model, developed in partnership with Yale University, has so far led to the formulation of a novel hypothesis about the behavior of cancer cells, potentially opening the door to new therapies to fight cancer.

Google noted in its blog post that the announcement marks a milestone for AI in science, and “reveals a promising new pathway for developing therapies to fight cancer.”

It’s not the only time that Google has experimented with AI as a means of scientific discovery. The company has been using AI in scientific research for years, with efforts including protein structure prediction, predicting genetic mutations and materials discovery using deep learning. 

And on Thursday, Google DeepMind revealed a partnership with Commonwealth Fusion Systems to develop fusion as a clean energy source, Axios reported (The outcome of these efforts does, however, stand to benefit its AI development).  

Tech figureheads like OpenAI’s Sam Altman and SoftBank’s Masayoshi Son have repeatedly emphasized that AI will lead to the evolution of humanity, especially as these models become more powerful. Google’s efforts represent a tangible step in that direction, signalling that AI can, in fact, be used for more than writing emails or automating tedious tasks. 

However, OpenAI might be coming for Google’s lab coat: The company reportedly hired Alex Lupsasca, a black hole theoretical physicist, marking the first hire of the OpenAI for Science initiative, according to Axios.

TOGETHER WITH KORE.AI

Forrester Names Kore.ai A Leader In Cognitive Search

In the era of AI and human collaboration, enterprise search has outgrown its original role as a simple lookup tool. It’s now the intelligence layer that taps into true knowledge scattered across enterprise systems that shapes decisions and drives action – which makes the fact that Forrester has named Kore.ai one of their leaders in the cognitive search space such big news.

In their latest report (which you can access for free right here), Forrester breaks down how cognitive search is set to become the brains of accurate agentic AI, the critical factors leaders should examine before picking their search platform, and what sets Kore.ai apart as a leader in the category.  

Access the latest Forrester Wave report on Cognitive Search Platforms right here.

CULTURE

Pinterest dials down slop

People are getting tired of AI slop. Pinterest is trying to limit it. 

On Thursday, the social media firm added the ability to restrict AI-generated content in certain categories, aiming to “dial down the AI and add more of a human touch,” it said in a blog post. Pinterest will also make its labels for AI content more noticeable. 

Pinterest’s Chief Technology Officer Matt Madrigal said in a statement that the company is “striking the right balance between human creativity and AI innovation, and ensuring every feed truly reflects what inspires them most.”

The move marks Pinterest’s second attempt at regulating slop after backlash that AI-generated content has taken over the platform. It signals a sharp rise in AI-generated content across the board as platforms like OpenAI’s Sora and Meta’s Vibes make their way to consumers’ hands.  Pinterest noted in its press release that AI content now takes up 57% of all online material. 

The growth in AI-generated content, combined with the evolving capabilities of these models and the proliferation of watermark removers, has made it increasingly difficult for the average user to distinguish between real and fake. 

For example, of the nearly 40,000 responses that The Deep View has garnered for our daily “Real or AI” game since July 21, roughly 51% of users are able to identify which image is authentic. (To test your own AI detection skills, check out the game below.) 

It highlights that these platforms may have to think of better ways to monitor and label this content, as these models create content that gets closer to reality every day.

LINKS

• DoorDash, Waymo partners on deliveries in Phoenix

• AI gaming startup General Intuition lands $134 million seed round 

• Uber to turn drivers into data labelers for extra cash

• Oracle to help OpenAI handle export controls for AI chips

• OpenAI is working on “sign in with ChatGPT” feature for websites

• Amazon’s Ring partners with Flock, an AI camera firm that works with ICE

• Google Veo 3.1: The latest version of Google’s video generation model, with more control, longer videos and native audio. 

• Simulon: Studio-quality visual effects development available on mobile. 

• CTO.New: This tool plans and ships code for free using agentic AI. 

• Zavo: AI-powered point of sale for brick-and-mortar, used by more than 400 businesses. 

• Amp Free: Amp has made its agentic coding capabilities free to use.

James, Tech company, Austin:

• We needed someone who could take our data infrastructure from “messy but working” to “smart and scalable.” Through Athyna’s platform, we matched with a remote ML engineer in days—not months. Six months later, latency dropped by 40% and deployment speed doubled. Stack: AWS, Spark, Databricks, Terraform, MLflow. Saved 68% on hiring costs.

Elena, E-commerce company, Miami:

• Finding the right talent for computer vision was like searching for a needle in a haystack—until Athyna made it simple. The specialist we hired streamlined our QA process and boosted visual accuracy to 97%. Stack: PyTorch, OpenVINO, Azure IoT, Docker, Kubernetes. Reduced time-to-hire from 60 to 9 days.

(sponsored)

GAMES

POLL RESULTS

What's your biggest concern about the neocloud boom?

• Environmental impact and energy consumption (47%)

• Potential market bubble or overvaluation (24%)

• Unequal access to AI computing resources (9%)

• Security and reliability concerns (18%)

• No major concerns (2%)