The Daily Pulse: AI's Reckoning Has Arrived

The Capital Surge: Trillions in Motion

OpenAI has closed a deal to raise $122 billion at an $852 billion valuation, its largest funding round to date as the company is expected to hit the public markets this year. The significance here transcends the dollar figure. OpenAI is generating $2 billion in revenue per month and taking a shot at competitors: "At this stage, we are growing revenue four times faster than the companies who defined the Internet and mobile eras, including Alphabet and Meta."

This isn't just a funding round. Amazon.com Inc. agreed to invest $50 billion in the round, while Nvidia Corp. and SoftBank Group Corp. each put in $30 billion. The structuring reveals the era we're entering: A large portion of Amazon's investment — $35 billion — is contingent on OpenAI going public or reaching the technological milestone of artificial general intelligence.

But here's what the hype misses: The eye-watering level of funding came in higher than originally projected, reflecting the surging costs of computing power and arriving amid lingering questions about whether OpenAI and other AI companies can generate sufficient revenue to cover expenses.

Across Big Tech, the capex arms race is unmistakable. At CES 2026, Samsung co-CEO TM Roh said that the company will double the number of Gemini-powered mobile devices it makes this year, bringing the total to 800 million. Samsung is adding these features to TVs and home appliances as well.

The Model Wars: AI Has Crossed a Threshold

On GDPval, which tests agents' abilities to produce well-specified knowledge work across 44 occupations, GPT‑5.4 achieves a new state of the art, matching or exceeding industry professionals in 83.0% of comparisons, compared to 70.9% for GPT‑5.2. This is the headline. The reality underneath is more consequential.

GPT-5.4 is OpenAI's first general model with native computer use. Agents can use screenshots, mouse, and keyboard input to control websites and software, handling complex tasks on their own. On OSWorld-Verified, which measures a model's ability to navigate a desktop environment through screenshots and keyboard/mouse actions, GPT‑5.4 achieves a state-of-the-art 75.0% success rate, far exceeding GPT‑5.2's 47.3%, and surpassing human performance at 72.4%.

The gap between capability and safety, however, remains stark. Since the publication of the 2025 International AI Safety Report, the number of companies publishing Frontier AI Safety Frameworks has more than doubled, and researchers have refined techniques for training safer models and detecting AI-generated content. However, significant gaps remain: sophisticated attackers can often bypass current defences, and the real-world effectiveness of many safeguards is uncertain.

The Human Cost: The AI Transition Begins

On March 31st, the reality of AI's economic impact became unmistakable. Tech giant Oracle has laid off thousands of its workforce in massive job cuts that could reportedly impact 30,000 employees globally amid the company's push to fund more artificial intelligence infrastructure. Employees in the U.S., India, Canada, Mexico, and other countries began receiving job termination emails from "Oracle Leadership" at about 6 am local time Tuesday.

This wasn't a surprise to analysts. Cutting 20,000 to 30,000 employees could lead to $8 billion to $10 billion in incremental free cash flow, TD Cowen analysts wrote in a January note. The logic is cold: The layoffs are closely linked to Oracle's aggressive push into artificial intelligence infrastructure. The company is investing heavily in data centres and cloud capacity, including large-scale partnerships with OpenAI.

Oracle isn't alone, and this isn't temporary disruption—it's structural reallocation. More than 70 tech companies have cut around 40,480 jobs so far this year, per Layoffs.fyi, as companies increasingly reallocate resources toward AI.

Security Unravels: Even the "Safe" Companies Can't Protect IP

On March 31, Anthropic accidentally exposed the full source code of Claude Code (its flagship terminal-based AI coding agent) through a 59.8 MB JavaScript source map (.map) file bundled in the public npm package @anthropic-ai/claude-code version 2.1.88.

The leaked file contained approximately 513,000 lines of unobfuscated TypeScript across 1,906 files, revealing the complete client-side agent harness. The scale and speed of replication: Within hours, the codebase was downloaded from Anthropic's own Cloudflare R2 bucket, mirrored to GitHub, and forked tens of thousands of times. Thousands of developers, researchers, and threat actors are actively analyzing, forking, porting to Rust/Python and redistributing it.

The cleanup was bungled. According to GitHub's records, the notice was executed against some 8,100 repositories — including legitimate forks of Anthropic's own publicly released Claude Code repository.

What's particularly damaging: the leak revealed defensive mechanisms Anthropic built specifically to prevent competitors from stealing its capabilities. In claude.ts (lines 301-313), a flag called ANTI_DISTILLATION_CC, when enabled, sends anti_distillation: ['fake_tools'] in API requests. This tells the server to inject decoy tool definitions into the system prompt. The idea: if a competitor is recording Claude Code's API traffic to train their own model, the fake tool definitions corrupt that training data.

Safety Researchers Are Leaving

The gap between capability advancement and safety governance isn't just a data point—it's becoming a crisis of confidence. Since the release of the inaugural International AI Safety Report a year ago, we have seen significant leaps in model capabilities, but also in their potential risks, and the gap between the pace of technological advancement and our ability to implement effective safeguards remains a critical challenge.

Led by Turing Award winner Yoshua Bengio and authored by over 100 AI experts, the report is backed by over 30 countries and international organisations. It represents the largest global collaboration on AI safety to date. But even the authoritative assessment reveals the mismatch: While the report notes that industry commitments to safety have expanded over the past year, efforts to mitigate risk are lagging. The capability of LLMs to aid hackers has increased far faster, than our ability to detect and block their use in cyberattacks. "Unfortunately, the pace of advances is still much greater than the pace of [progress in] how we can manage those risks and mitigate them."

What This Means: A Structural Shift, Not a Cycle

We're not in an AI hype phase anymore. We're watching capital reallocation in real time—from people to chips, from services to infrastructure, from distributed work to centralized compute.

Samsung is entering 2026 with an ambitious goal: putting Google's Gemini AI into the hands of 800 million users. During an interview at CES 2026, Samsung's co-CEO TM Roh confirmed that the company plans to double its current AI-enabled device footprint. At the end of last year, it stood at roughly 400 million units.

The question isn't whether AI will displace workers. That's already happening. The question is whether the productivity gains, revenue growth, and new capabilities justify the human and economic costs. For now, the evidence is mixed: Developers can even configure the model's safety behavior to suit different levels of risk tolerance by specifying custom confirmation policies. This is progress for technical capability. It's also a way to quietly shift responsibility for safety decisions onto users and developers.

The next 12 hours will bring more stories. But the pattern is set: capital flows to compute, safety lags capability, workers are replaced, and the companies making this transition win. That's not a prediction. That's today's news.

Daily Pulse: The Great AI Reallocation—Infrastructure Wins, Workers Lose, and Compute Becomes Power

The Daily Pulse: AI's Reckoning Has Arrived

The Capital Surge: Trillions in Motion

The Model Wars: AI Has Crossed a Threshold

The Human Cost: The AI Transition Begins

Security Unravels: Even the "Safe" Companies Can't Protect IP

Safety Researchers Are Leaving

What This Means: A Structural Shift, Not a Cycle

Complete Sources & Further Reading

Daily Pulse: The Great AI Reallocation—Infrastructure Wins, Workers Lose, and Compute Becomes Power

The Daily Pulse: AI's Reckoning Has Arrived

The Capital Surge: Trillions in Motion

The Model Wars: AI Has Crossed a Threshold

The Human Cost: The AI Transition Begins

Security Unravels: Even the "Safe" Companies Can't Protect IP

Safety Researchers Are Leaving

What This Means: A Structural Shift, Not a Cycle

Complete Sources & Further Reading

Daily Pulse: The $500B Bet—How Tech Giants Are Dividing the Future

Daily Pulse: The Trillion-Dollar Consolidation—How 12 Hours Redefined AI's Future

Daily Pulse: The AI Reckoning Arrives—$650B in Bets, Broken Models, and the Fragmentation Wars