OpenAI Usage Soars

In the early days of computing, innovation often came from small labs turning ideas into revolutions.

Today, OpenAI stands as a prime example of that relentless drive.

The San Francisco tech company now boasts 400 million weekly active users—a 33% jump from December—underscoring its rapid growth and rising influence.

This surge is not merely a matter of consumer adoption.

OpenAI’s Chief Operating Officer, Brad Lightcap, highlighted that ChatGPT’s increasing utility and familiarity drive word-of-mouth growth. “People hear about it, see the value, and then discover their own use cases,” he explained. In simple terms, what began as an experimental tool is quickly becoming indispensable.

Short bursts of enthusiasm ripple across social and professional networks, fueling both personal and enterprise engagement.

The company’s enterprise business tells a parallel story.

With 2 million paying enterprise users—doubling since September—employees using ChatGPT on a personal level have become powerful advocates within their workplaces.

This organic push is propelling developer traffic and prompting major companies like Uber, Morgan Stanley, Moderna, and T-Mobile to integrate OpenAI’s technology into their own platforms.

The analogy to early cloud services is not lost on Lightcap; he envisions AI as the next essential infrastructure for modern business.

Yet, not all eyes are friendly.

Chinese competitor DeepSeek has stirred controversy and concern among investors, even triggering a significant dip in tech market valuations.

OpenAI’s response was swift, accusing DeepSeek of misappropriating its models—a move that reinforces the competitive intensity in the AI space.

Adding to the tension, legal battles and high-stakes bids involving figures like Elon Musk are reshaping the company.

Amidst billions in investments and market pressures, OpenAI remains resolute: it is not for sale.

As the debate over innovation versus consolidation continues, one wonders: in the rush toward an AI-driven future, will competition fuel creativity or curb its potential?

Scaling AI: Half-Trillion Dollar Bet Hits a Wall

Massive investments in AI are increasingly at risk as the “scaling hypothesis” shows signs of faltering.

Investors have funneled on the order of half a trillion dollars into the belief that simply adding more data and GPUs will eventually unlock artificial general intelligence.

This premise, once boldly declared by researchers like DeepMind’s Nando de Freitas and celebrated in t-shirt slogans proclaiming “scale is all you need,” now faces mounting skepticism.

The scaling hypothesis rests on the notion that enlarging models will automatically yield superintelligent machines capable of outperforming humans in every cognitive task.

A landmark 2020 OpenAI paper even suggested that performance improvements could be predicted by simple equations—a power law linking model size, dataset size, and compute.

For a while, those predictions held water. But critics have long warned that these measures merely capture surface-level performance. They miss the deeper kind of comprehension that genuine intelligence requires.

As early as March 2022, voices in the field argued that deep learning was “hitting a wall.”

Today, those warnings seem increasingly prescient.

Recent industry leaks—including investor Marc Andreessen’s comment about models nearing a capability ceiling and Satya Nadella’s admission that scaling laws are just empirical observations—add fuel to the fire.

Companies are beginning to pivot. Heavy hitters like Klarna have abandoned their all-in AI strategies, and projects such as Humane AI Pin have been scrapped.

Even OpenAI admits that its next-generation models will not emerge solely from scaling up.

Errors, hallucinations, and unreliable reasoning persist despite larger, more expensive models.

As a result, the current playbook—used by everyone from OpenAI to Chinese firms like DeepSeek—may be doomed to deliver only incremental improvements.

The challenge remains: can the AI industry break free from a cycle of price wars and diminishing returns to truly innovate, or will these investments prove unsustainable in the long run?

Gemini Breaks Free on iOS

Google is making a bold move by pulling its AI assistant, Gemini, from the main Google app on iOS devices.

Instead, users will now be directed to a standalone Gemini app.

This shift is designed to position Gemini as a direct competitor to consumer-facing AI chatbots such as ChatGPT, Claude, and Perplexity.

The decision was communicated via an email alerting users that “Gemini is no longer available in the Google app.” In the message, Google urged iOS customers to download the dedicated Gemini app from the App Store.

The app, which launched late last year, now serves as the exclusive home for Gemini on iOS. It offers a range of features including voice conversations through Gemini Live, integration with Google services like Search, YouTube, Maps, and Gmail, as well as capabilities to answer questions, plan trips, generate images, and more.

Some users will interact with Gemini via text, others by voice or even using their device’s camera.

The versatility of interaction is a key selling point. However, the email also reminded users that Gemini can still produce mistakes, urging them to verify its responses—a caution that underscores the tool’s current limitations.

Additionally, Google is leveraging the standalone app to promote its paid subscription model.

Through the app, users can upgrade to the Google One AI Premium plan for access to Gemini Advanced. In the main Google app, any attempt to access Gemini now triggers a full-screen message stating “Gemini now has its own app” with a link to download it.

This move is a calculated risk. While the standalone app could allow Google to rapidly roll out new AI features and compete more directly in the market, it might also lead to reduced usage.

Millions of iOS users already have the Google app, and not all may be inclined to download another app.

Murati’s New AI Lab

Former OpenAI CTO Mira Murati has launched a new AI startup, Thinking Machines Lab, with a bold mission to bridge key gaps in frontier AI systems.

The startup aims to make AI more widely understood, customizable, and capable.

The news comes six months after Murati’s departure from OpenAI, and the company is already stirring interest.

Murati’s blog post revealed a leadership team composed of former colleagues and prominent figures from top AI labs.

Among them, OpenAI co-founder John Schulman has been named chief scientist, rejoining forces with Murati after his exit from Anthropic.

Barret Zoph, previously OpenAI’s vice president of research, takes the role of chief technology officer at Thinking Machines Lab. In total, the startup announced 29 employees, including at least seven ex-OpenAI veterans.

Jonathan Lachman, the former head of special projects at OpenAI, is one of the high-profile hires, along with researchers such as Lilian Weng, Luke Metz, Sam Shleifer, and Stephen Roller.

The firm’s mission is clear: to decentralize the knowledge of AI system training that is currently monopolized by elite research labs. “Knowledge of how these systems are trained is concentrated within the top research labs, limiting both the public discourse on AI and people’s abilities to use AI effectively,” the blog post explained.

Murati left OpenAI in September 2024 to “create the time and space to do my own exploration.”

Her decision appears to have sparked a wave of innovation, as the new startup is already in talks to raise over $100 million in funding, according to recent media reports.

The ambitious team also includes researchers from Meta, Google DeepMind, CharacterAI, and Mistral.

This diverse mix of talent signals that Thinking Machines Lab is set on challenging the status quo in AI research and product development.

In a fast-evolving industry, the real test will be whether fresh ideas can break through the limitations of current AI models and truly reshape the future of technology.