A.I. in the Workplace

OpenAI model escapes sandbox, exploits zero-day to breach Hugging Face

Google increases capex forecast again after cloud-driven quarterly beat

DeepSeek chats are showing up on Google… and that’s raising some serious privacy questions.

It turns out that conversations shared through DeepSeek’s “public link” feature can be indexed by search engines. That means anything from casual chats to sensitive work discussions could potentially be discovered by anyone online.

Important detail: this isn’t a hack or data breach. These were links users chose to share. But here’s the catch — once something is public, it can spread far beyond what you intended. And without an easy way to track or revoke those shared links, people may not even realize what’s still out there.

This situation highlights a bigger issue in the AI space:
We focus so much on how powerful these tools are… but not enough on how securely they handle our data.

As AI becomes part of everyday life, privacy design matters just as much as performance.

Would you trust AI platforms with sensitive information knowing this?

AMD Stock Gains on Expanding AI Push With 2GW Anthropic Deal and $5B Investment Plan

Advanced Micro Devices (AMD) and Anthropic have signed a major agreement to deploy up to 2 gigawatts of AMD’s Instinct MI450‑series GPUs using the company’s Helios rack‑scale systems. The first gigawatt is set to roll out in the first half of 2027. AMD also committed to invest up to $5 billion in Anthropic over a period of time.

With this deal, Anthropic will move to a full AMD stack — MI455X GPUs, EPYC “Venice” CPUs, Pensando networking, and ROCm software — a big step up from its current MI355X footprint.

The partnership goes deeper than hardware. Anthropic will use Claude to help AMD speed up software work, tune workloads for Instinct GPUs, and push faster updates across ROCm. AMD plans to use Claude across its engineering teams.

AMD CEO Lisa Su said gigawatt‑scale Helios deployment is a key move for next‑gen AI infrastructure. For Anthropic, the added compute comes at the right time. Claude demand has surged so fast that usage has been limited and outages have hit some customers.

The White House released a new report on Tuesday that could signal a fundamental shift in the government’s $200 billion annual research and development budget, The Wall Street Journal reports. Office of Science & Technology chief Michael Kratsios said the report positions the U.S. to "lead the AI-driven scientific revolution" by prioritizing individuals and AI over universities and institutes. While some academics and researchers on LinkedIn praised the document and parsed its details, others pointed out recent government budget cuts and funding reductions toward scientific groups and organizations.

🚨 The AI Boom Is Triggering a Massive CPU Shortage! 🚨

Think the AI craze is only about GPUs like Nvidia’s? Think again. The massive surge in AI data center construction is now driving a heavy global demand for standard server CPUs.

According to a new Reuters report, tech giants Intel and AMD are negotiating long-term supply commitments (1–2+ years) with Chinese server clients as server processor prices skyrocket.

Here’s what’s happening: 🔹 Prices Skyrocketing: Server CPU prices in China are up over 40% since the start of the year, with month-on-month jumps topping 10%. 🔹 Long Lead Times: Delivery wait times for some Intel Xeon processors have reached up to 6 months. 🔹 Why Now? AI data centers don't just run on GPUs—they need massive armies of CPUs to manage networking, storage, and inference workloads.

As suppliers gain massive leverage, Chinese cloud providers are scrambling to secure supply despite ongoing trade restrictions.

The tech world is on edge this week after OpenAI blamed two of its AI models for breaching Hugging Face's internal systems. Experts say the incident shows that even elaborate containment measures for AI models may not be adequate to protect against potentially harmful cyberattacks. Tech and AI firms have faced heightened scrutiny since June, when the federal government forced Anthropic to temporarily restrict access to Mythos 5 and Fable 5 over national security concerns. Since then, Google, OpenAI and other firms have vowed to improve security capabilities.

 Alphabet (GOOGL.O)

, opens new ta on Wednesday reported its best-ever quarter of growth for its cloud ​computing division, but faced investor scrutiny after concerns about continued delays to its flagship AI model were aggravated by a $15 billion increase in capital spending ‌plans for 2026.

The search giant now expects to spend between $195 billion and $205 billion in capital expenditures, its finance chief Anat Ashkenazi said on a conference call with analysts. The company said last quarter that it planned to spend between $180 billion and $190 billion this year.

Shares of the company were down about 3% in extended trading. The stock was initially volatile but mostly flat, but dipped after Ashkenazi announced the capex update.

"We have increased ​our capacity quite significantly over the past three years. The demand still outpaces that investment," Ashkenazi said in justifying the increased spending plans. She added that faster-than-expected delivery ​of capacity contributed to the hike.

Revenue at Google Cloud rose 82% to $24.8 billion during the quarter ended June, driven by strong demand from ⁠AI-hungry enterprises worldwide. Analysts on average expected a 64% increase, according to data compiled by LSEG.

Advertising revenue came in at $81.6 billion compared to estimates of $81.1 billion. Total revenue for the ​quarter was $119.8 billion, beating the consensus estimate of $116.9 billion.

But adjusted profit per share of $2.85 fell slightly short of Wall Street projections of $2.89. And the company reported negative free cash flow for ​the first time in its history, burning $5.9 billion this quarter.

"After a negative cash flow quarter, the new raise in capex does not sit well for Alphabet," said Thomas Monteiro, senior analyst at Investing.com. "The market's most reliable cash generators are now spending more than they bring in. As long as revenue keeps accelerating, investors will tolerate it. But capital has a real cost again, and the room for error is shrinking ​every quarter."

MODEL UNCERTAINTY TEMPERS CLOUD GAINS

While Google Cloud has made Alphabet a big beneficiary of the AI boom, the company's own AI efforts have lost some steam this year after ​it delayed the June launch of its next flagship model, Gemini 3.5 Pro.

That has left Google trailing in the AI coding tools market and fueled concerns on Wall Street, especially as Anthropic and OpenAI have ‌consistently rolled ⁠out enterprise-focused upgrades, and Chinese open-source models have also gained strong traction.

During the question-and-answer portion of the earnings call, multiple analysts pressed CEO Sundar Pichai for clarity around whether Google could keep up with its rivals at the frontier of model development.

"There are many attributes on which we are still at the frontier. There are areas where we've acknowledged we need to improve; coding and agentic coding is an example of that," Pichai said on the call.

Pichai said that while Google continued to test Gemini 3.5 Pro, it had also started training Gemini 4 and ​was "applying a lot of our compute and effort ​in that direction" in order to ⁠remain competitive.

"We are both very committed and very confident of being at the frontier for the next generation," he said.

Meanwhile, Google began recognizing revenue from direct sales of its TPU chips, which compete with Nvidia's (NVDA.O), opens new tab GPUs, for the first time in the second quarter, though the vast ​majority of revenue from business agreements would come through next year, Ashkenazi said.

The third-largest cloud services provider behind Amazon Web Services (AMZN.O), opens new tab and Microsoft (MSFT.O), opens new tab, Google has seen ⁠demand surge as companies race to secure the cloud capacity needed to develop, train and run AI models, helping it land major deals with firms, including Anthropic.

Big Tech is expected to spend well over $700 billion this year primarily on AI, while Morgan Stanley has pegged the estimated spend at more than $1 trillion for the next year.

Ashkenazi reaffirmed on the call that Alphabet planned another significant ⁠increase to capex ​in 2027.

Alphabet shares have been among the best performers in the "Magnificent 7" group of stocks this year so ​far, rising more than 9% through last close. But concerns over the Gemini delays, some high-profile executive departures, and regulatory pressures have dragged the stock about 9% lower since the end of April.

Tesla just dropped their Q2 2026 financial results and the numbers are not pretty. Despite a revenue beat, margins are down. Lower average selling price for the company's cars, and a continued if expected decline in automotive regulatory credits figured in, and CapEx / OpEx is just soaring.

Meanwhile, the company's update on Optimus so far is just not exciting. They're building factory lines out but said the first "builds" of the humanoid robot won't be deployed to pilot customers, just used for training and timelines are not well-defined.

Meanwhile, Tesla said they are "ramping unsupervised" Robotaxi rides, i.e. actually driverless rides in every market where they are trying these. Far cry from Waymo's lead in the US and Baidu's with Apollo Go in China.

Rivals Microsoft and Amazon are set to ​report their quarterly earnings next week.