Google might launch an AI product with one name during its experimental phase, but it will eventually tie it all to Gemini. In the latest example of this trend, the company is renaming its AI-powered research product NotebookLM to Gemini Notebook. The company is also adding features to make the tool more interactive by infusing coding execution for data analysis.

The company first showed off NotebookLM during Google IO in 2023 as Project Tailwind, and since then, it has made it into a product used by 30 million people and over 600,000 organizations. In the last three years, the company has added capabilities, like interactive podcast generation, curated notebooks, video overviews, support for more file types, and an enterprise plan.

Because of NotebookLM, other companies and startups have added capabilities for podcast generation from source material and research tools.

Along with renaming, Google is rolling out a new update that makes each notebook its own secure container, in which users can generate code to make outputs interactive. It noted that with code execution ability, users can tap into multiple sources and create complex data analysis directly within the tool.

The company said the update is available to Google AI Ultra paid plan users, along with Workspace business customers with AI Ultra Access and AI Expanded Access. Pro users will get access to this feature in the coming weeks.

Google said that users can already look at their notebooks within the Gemini app, and soon, they will be able to access them through AI Mode in search.

Enterprises looking to move more of their agentic AI workloads to open weights models they can customize, control and run on-premises or in virtual private clouds have a strong new contender to consider.

Today, Thinking Machines—the highly capitalized American AI startup founded by former OpenAI CTO Mira Murati—released Inkling, its first major language model under an enterprise-friendly Apache 2.0 open source license, and it boasts high, if sub state-of-the-art, performance for open weights models on third-party benchmarks, specifically software engineering (77.6% on SWE-bench Verified, where it beats fellow U.S. open rival Nvidia Nemotron 3’s 71.9%) and voice understanding (91.4% on VoiceBench compared to 94.4% for Gemini 3.1 Pro on high reasoning effort).

Another differentiator: Thinking Machines notes that Inkling was designed “to answer directly on topics that may be subject to censorship,” offering enterprises concerned about factual outputs, irrespective of controversy or sensitivity, a more trustworthy option.

Coming in at 975 billion total parameters, Inkling is a natively multimodal, open-weights Mixture-of-Experts (MoE) system capable of reasoning across text, images, and audio. The weights are already available on Hugging Face and the company’s own model training application programming interface (API), Tinker.

Designed to balance cost against performance through a novel “controllable thinking effort” mechanism, the model represents a significant departure from the black-box scaling strategies of frontier competitors.

Alongside the flagship model, Thinking Machines also announced a preview of Inkling-Small, a lighter 276-billion-parameter alternative optimized for workloads where low latency and cost are paramount.

Benchmarks Show a Powerful, High-End, Sub State-of-the-Art Model

While Inkling is a formidable multimodal engine, it lands in a fiercely competitive 2026 open-weight landscape characterized by highly specialized MoE architectures. Rather than attempting to dominate every leaderboard, Thinking Machines explicitly designed Inkling—with 975 billion total and 41 billion active parameters—as a broad, balanced generalist.

For example, it comes in near the middle high-end of benchmark performance 1257 on Design Arena’s Agentic Web Dev leaderboard measuring human scores of frontend web design.

But China’s leading AI labs have produced models with elite reasoning and coding capabilities, posing a stiff challenge to Inkling’s generalist approach and ultimately outperforming it on general and coding benchmarks.

• GLM 5.2: Widely considered the top open-weight reasoning model available in the benchmark set, GLM 5.2 outperforms Inkling on pure coding, agentic, and complex reasoning tasks. It scores 62.1% on SWEBench Pro (Public) compared to Inkling’s 54.3%, and a massive 82.7 on Terminal Bench 2.1 against Inkling’s 63.8. GLM 5.2 also holds the edge in text-only reasoning, scoring 40.1% on HLE (text only) versus Inkling’s 30.0%.

• DeepSeek V4 Pro: DeepSeek maintains an edge in several strict coding and factuality domains, beating Inkling on SWEBench Verified (80.6% vs. 77.6%) and SimpleQA Verified (57.0% vs. 43.9%). However, Inkling successfully overtakes DeepSeek V4 Pro in mathematical problem-solving, achieving 97.1% on AIME 2026 compared to DeepSeek’s 96.7%.

• Kimi K2.6: This model outpaces Inkling across multiple technical benchmarks, delivering higher scores on GPQA Diamond (91.1% vs. 87.9%), BrowseComp (83.2% vs. 77.1%), and HLE with tools (54.0% vs. 46.0%). Yet Inkling proves more resilient on general chat instruction following, scoring 79.8% on IFBench compared to Kimi K2.6’s 76.0%.

Against its primary U.S.-based open-weight competition, Inkling demonstrates strong parity and frequent superiority.

• Nemotron 3 Ultra: Inkling consistently outperforms this U.S. rival across reasoning and coding. Inkling posts 97.1% on AIME 2026 and 77.6% on SWEBench Verified, beating Nemotron’s 94.2% and 70.7%, respectively. Furthermore, Inkling significantly leads in agentic workflows, scoring 74.1% on MCP Atlas against Nemotron’s 44.7%.

When compared to closed-source juggernauts like Claude Fable 5, GPT 5.6 Sol, and Gemini 3.1 Pro, Inkling trails in peak reasoning and software engineering autonomy, but remains highly competitive in multimodality.

• Coding and Reasoning: Closed models maintain a commanding lead. Claude Fable 5 (max) hits 95.0% on SWEBench Verified and 53.3% on HLE (text only), far outpacing Inkling’s 77.6% and 30.0%. GPT 5.6 Sol dominates Terminal Bench 2.1 with an 89.5, easily clearing Inkling’s 63.8.

• Native Multimodality: Inkling’s native visual and audio capabilities hold their own. On the MMMU Pro (Standard 10) vision benchmark, Inkling’s 73.3% is competitive, though trailing Claude Fable 5’s 84.2% and GPT 5.6 Sol’s 83.0%. In audio processing, Inkling scores a highly respectable 77.2% on MMAU, keeping it within striking distance of Gemini 3.1 Pro’s 82.5%.

If an enterprise workflow demands elite software engineering autonomy or the highest bounds of text-only reasoning, models like GLM 5.2 or proprietary systems like Claude Fable 5 maintain the edge.

However, Inkling carves out a unique and highly defensible position: it is the most capable open-weight foundation model that natively fuses text, vision, and audio, while simultaneously offering developers direct programmatic control over the cost-to-performance ratio.

The Shift from Static Reasoning to Controllable Thinking

Rather than attempting to build a singular “god model” optimized strictly for state-of-the-art benchmark domination, Thinking Machines engineered Inkling for adaptability and efficiency in real-world workflows.

The standout feature of this release is Inkling’s “controllable thinking effort.” Developers can programmatically adjust the model’s reasoning budget—scaling from 0.2 to 0.99—to dictate how hard the AI should “think” before generating an output.

As the company noted, “Inkling’s continuous thinking effort lets you pick your point on the cost/performance curve—reaching the same score with a fraction of the tokens”.

In practical terms, this allows enterprises to deploy Inkling with lower token expenditure for simpler tasks, while cranking up the compute overhead for complex, multi-step reasoning challenges. However, by keeping the thinking effort lower and generating fewer tokens, the cost-conscious enterprise can achieve high quality results and performance on simple tasks while spending less money, or, in the case of those running models locally, less costs on energy and compute resources.

During the model’s large-scale reinforcement learning (RL) training over 30 million rollouts, researchers observed an emergent phenomenon they called “chain of thought condensation”. Over time, Inkling naturally learned to compress its internal reasoning steps—dropping grammatical overhead and connectives—while reaching the same accurate conclusions, resulting in drastically reduced latency.

Epistemics and Censorship Resistance

A notable element of Thinking Machines’ release is its explicit focus on the model’s epistemics—specifically its calibration, instruction following, and resistance to censorship.

In an ecosystem where open-weight models adopt either overly restrictive safety guardrails or echo state-aligned ideological talking points, Inkling was intentionally trained to answer directly on politically sensitive or heavily censored topics.

To validate this approach, Thinking Machines submitted Inkling to the Propaganda and Censorship Eval developed by AI startup Cognition. According to the published findings, Inkling demonstrated “strong patterns of censorship non-compliance,” effectively resisting ideological capture or boilerplate refusals when presented with sensitive subjects.

Despite its resistance to censorship, the model maintains a robust defense against genuinely malicious, dangerous, or illegal queries. On the StrongREJECT benchmark—which tests responses to unambiguous harmful requests—Inkling scored 98.6%, placing it in line with strict frontier safety standards. Furthermore, on the FORTRESS benchmark, Inkling successfully navigated the line between safety and over-refusal: it achieved a 78.0% refusal rate on adversarial queries (such as those involving weapons, cyberattacks, or violence) while maintaining a 95.9% compliance rate on benign, look-alike queries.

Thinking Machines noted that typical open-weight vulnerabilities remain within the architecture. Internal safety evaluations revealed an “occasional tendency to comply with role-play and indirectly framed prompts concerning harmful topics”. The company advised enterprise developers to treat the model’s built-in refusals as just one layer of security, recommending the downstream deployment of external moderation tools—such as Llama Guard—to filter adversarial jailbreaks and enforce use-case-specific safety policies at the application level.

Under the Hood: Architecture and Multimodality

Inkling’s scale is staggering, yet sparse. The MoE architecture features 975 billion total parameters, but only 41 billion parameters are active during any given token generation. It supports a massive context window of 1 million tokens and diverges from typical transformer models by using relative positional embeddings instead of the industry-standard Rotary Positional Embedding (RoPE).

True to the company’s foundational vision, Inkling was trained from scratch to be natively multimodal. Unlike models that rely on bolted-on external encoders, Inkling uses an encoder-free early fusion approach. It directly ingests audio as discrete dMel spectrograms and visual data as 40×40 pixel patches via a hierarchical multi-layer perceptron (hMLP), projecting all modalities into a shared hidden space.

Licensing: True Open-Source for the Enterprise

For enterprise IT teams and developers, the most disruptive aspect of Inkling may be its licensing. Inkling is released under the permissive Apache 2.0 license.

In an ecosystem where many so-called “open” models from Western labs are tethered to dual-use commercial licenses, acceptable use restrictions, or revenue caps, an Apache 2.0 designation makes Inkling a true open-source foundation. This gives developers the legal freedom to download, modify, integrate, and commercialize the model weights entirely royalty-free.

The model is readily deployable across major open-source inference libraries—including SGLang, vLLM, TokenSpeed, and llama.cpp—and comes with a native NVFP4 quantized checkpoint optimized for NVIDIA Blackwell systems.

Community Reactions: The Engineering Feat

The AI community’s response has been swift, praising both the model’s openness and the underlying engineering execution.

In a post on X, Thinking Machines co-founder John Schulman reflected on the rapid development cycle: “Inkling is out today, with open weights and in Tinker. It’s been fun to watch this one come together: pretraining began last winter, and starting in mid-January a small team built up the coding, reasoning, and agentic training from there. We learned a lot building it, and I hope people find good uses for it.”

Horace He, a researcher at Thinking Machines (previously from PyTorch), underscored the difficulty of the task in another post on X: “It truly takes a village to release a model, perhaps especially an open weights model. Actually doing the entire process from scratch, from data to pretraining to posttraining to actual release, gives a lot of appreciation for anyone who does it!”

The broader open-source ecosystem has also embraced the technical integrations. Lysandre Debut, the Chief Open-Source Officer at Hugging Face, shared his enthusiasm regarding the model’s optimization in his own X post: “One thing I find quite striking is how much easier accelerating models has become… We replaced the model’s causal Conv1D with the `causal-conv1d` kernel. One line changed, +4% tokens per second. We then replaced its attention implementation with FlashAttention-4. Another single change, another +11%. That’s a total throughput improvement of about 15%, without changing the model architecture or retraining anything.”

Tiezhen Wang, an ecosystem growth expert and ex-Googler, celebrated the release as a massive win for the open-source community, listing the model’s impressive specifications on X, highlighting its “975B total, 41B active” size, “Native MTP support,” and the highly coveted “Apache 2.0 license.”

Background: The Road to Inkling

To understand the significance of Inkling, one has to look back at the rapid trajectory of Thinking Machines over the past 18 months.

When Mira Murati departed OpenAI in late 2024 to found Thinking Machines alongside industry veterans like John Schulman and Barret Zoph, the stated goal was to pivot away from building isolated autonomous agents. Instead, the company aimed to build flexible, multimodal systems designed for genuine human-AI collaboration and open science.

By July 2025, the startup had secured a historic $2 billion seed round led by Andreessen Horowitz at a $12 billion valuation. At the time, Murati promised the impending release of a product with a “significant open source component” to empower researchers and startups.

The company’s philosophy began coming into sharper focus in October 2025 with the launch of Tinker, a Python-based API for large language model fine-tuning that gave researchers granular control over training pipelines without the friction of distributed compute management.

That same month, Thinking Machines researcher Rafael Rafailov delivered a provocative critique of the AI industry at TED AI. He argued that the current trajectory of simply throwing more compute at models was fundamentally flawed, noting that today’s systems take shortcuts—like wrapping code in try/except blocks—because they are trained strictly for task completion rather than genuine learning.

Rafailov posited that the first artificial superintelligence would not be a “god model,” but rather a “superhuman learner” capable of meta-learning and internalizing abstractions. Inkling’s architecture—specifically its controllable thinking effort and its ability to organically compress its chain of thought during RL—feels like the first tangible realization of Rafailov’s thesis.

In May 2026, the lab teased its technical prowess with the research preview of TML-Interaction-Small, a system that eliminated “turn-based” chat by processing inputs and outputs simultaneously in 200ms chunks. This “full-duplex” breakthrough proved the company could build highly responsive, natively multimodal models from scratch.

Now, with Inkling out in the wild, Thinking Machines has delivered on its foundational promises. By offering a massive, natively multimodal model under a true open-source license, they aren’t just giving developers a new tool—they are attempting to fundamentally rewrite the economics and accessibility of frontier AI development.

from the do-not-pass-go,-do-not-collect-$200 dept

Not long after twelve states sued Paramount claiming its $111 billion merger with Warner Brothers would harm market competition, the Writers Guild of America (WGA) filed their own lawsuit, warning that the massive debt load from the media industry’s latest megamerger will result in an ocean of layoffs for an already reeling U.S. entertainment industry.

The lawsuit notes that the current film industry is dominated by just five players: Disney (ABC), NBCUniversal (Comcast), Sony, Paramount (CBS), and Warner Brothers. Comcast recently restructured to make it easier to sell off its NBC and Universal properties, opening the door to a lot of very quick consolidation in addition to the speedy Skydance/Paramount/Warners merger.

“With fewer competitors, the merged Paramount-Warner Bros. entity would have both the ​incentive and the ability to lower costs by suppressing writers’ wages and reducing output. Writers will be paid less and ​have fewer employment opportunities,” the WGA complaint said.

Supreme Court precedent (for whatever that’s worth anymore) has long indicated that any merger
yielding a post-merger market share exceeding 30% (which this deal does) is presumptively anticompetitive. The WGA notes that muted competition will result not just in fewer jobs, but lower wages and fewer opportunities for creatives overall across both film and television.

“With fewer competitors, the merged Paramount-Warner Bros. entity would have both
the incentive and the ability to lower costs by suppressing writers’ wages and reducing output.
Writers will be paid less and have fewer employment opportunities,” the lawsuit states.

While Paramount would like to pretend this is a debate, and most U.S. press outlets bury the lede, U.S. history is vividly clear on the harms created by media consolidation. That was most recently personified by AT&T’s disastrous acquisitions of DirecTV and Time Warner, which resulted in upward of 50,000 layoffs, higher prices, worse service, and no shortage of shuttered creative projects.

The rushed acquisitions of both CBS/Paramount and Warner Brothers — all so Larry Ellison’s son can play media mogul — have created a particularly heavy debt load of $79 billion. Such debt is always paid for by consumers and labor, often in more ways than one.

Paramount has promised to release 30 theatrical releases per year and to keep them in exclusively for theaters for 45 days, but as I’ve long made clear, pre-merger promises are utterly worthless. Especially in a country dead set on steadily lobotomizing its public interest regulators. As we’ve seen with consolidation in sectors like wireless, America’s favorite pastime is pretending to ignore the harms of pointless mergers.

This is a pretty clear example of the kind of consolidation that should be blocked for the benefit of labor, markets, and consumers, but despite a lot of rambling pretense about a love of free market competition and entrepreneurial spirit, America consistently fails to walk the talk on antitrust, the impact of which is abundant and getting exponentially worse under pay-to-play Trumpism.

Filed Under: antitrust, consolidation, film, hollywood, jobs, larry ellison, layoffs, media, mergers, movies

Companies: paramount, warner bros., writers guild

SCIENCE

Blue Origin and SpaceX get their turn to prove they can dock, loiter, and not blow up the launch pad

NASA has given an update on the Artemis III mission and, while sticking with an optimistic 2028 landing target for Artemis IV, offered a glimpse into just how much development work remains to be done at Blue Origin and SpaceX. 

Artemis III has been compared to Apollo 9, which tested the Apollo Lunar Module in Earth orbit, yet neither SpaceX nor Blue Origin is flying anything as close to the lunar landers. 

Blue Origin’s test lander will be based on the company’s current Mark 2 crew lander architecture, incorporating the major avionics, flight software, life support, and crew cabin. Orion, launched atop NASA’s SLS, will dock to the side of the Blue Origin spacecraft for crew transfer; two crew members in orange Orion survival suits can baord the test lander, with Orion’s software controlling the stack.

An instrumented lunar surface spacesuit mass simulator, similar to the “Moonikin” manikin that flew aboard Orion for Artemis I, will also ride along on the Blue Origin lander.

SpaceX’s test is considerably simpler – just a docking system mounted on the nose of a Starship. That requires Starship testing to have reached the orbital stage first, which is why NASA will be closely watching the upcoming Flight Test 13. Starship V3 is still flying suborbital until SpaceX proves it can reliably relight an engine for controlled re-entry.

Under the current plan, Blue Origin launches its lander into orbit first, where it can loiter for up to 30 days. Once it’s checked out, a crew launches aboard Orion to rendezvous and dock with it. After that’s complete, SpaceX launches its Starship test article to rendezvous and dock Orion in turn, though the crew won’t board Starship, just verify communications and interoperability. SpaceX’s vehicle will control that docked stack.

Notably, SpaceX’s docking capability was qualified in 2023, while Blue Origin only tested its pressurized docking system earlier this year.

Jeremy Parsons, Artemis program manager, stated, “Artemis III will be a highly choreographed dance with a demanding launch sequence across multiple launch pads and equally demanding mission operations for our ground and flight crews, making it one of the most complex and ambitious missions NASA has ever undertaken.”

He is not exaggerating. Apollo 9 needed a single Saturn V launch; Artemis III needs three – an SLS, whatever Blue Origin ultimately uses to launch its lander (the company is still rebuilding its launch pad after May’s explosion), and a Starship. The SLS has flown twice, including one lunar flyby. Starship has yet to reach orbit despite Elon Musk once claiming that uncrewed versions would be landing on Mars around now.

It’ll be an impressive feat if NASA can pull it off, even if SpaceX’s piece of the puzzle looks a lot simpler than Blue Origin’s. ®

Soon the internet will be full of AI agents acting on our behalf. Right now, there is no reliable way to tell who stands behind any of them. Vint Cerf, one of the people who built the internet, wants to fix that.

Cerf co-designed TCP/IP, the protocol that lets the internet’s independent systems talk to each other. He left Google last week after 20 years. Now he is joining the advisory council of Innovation Labs, a group building an open identity layer for AI agents, the company announced.

The missing layer

The problem is simple to state. Most AI agents today live inside one company’s systems. But firms want them roaming the open web, dealing directly with other agents. There is no shared way to prove who owns an agent, or who answers for what it does.

Innovation Labs is a division of Identity Digital, a firm that runs domain-name registries. Its idea, called DNSid, would give each agent a lasting identity tied to an existing domain name, backed by cryptographic proof. It has already submitted the design to the internet’s main standards body.

Why Cerf signed up

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

Cerf frames it as the internet’s next big architectural problem. The trigger, he told TechCrunch, is “the question of what authorities they have, where they have derived those authorities, who is accountable.”

He expects it to be messy. “It’s going to be a fascinating, and at the same time maybe even exasperating, period,” he said. Rival standards are already appearing. Cerf thinks none will win on politics, only on what works, as happened with TCP/IP.

Keeping it open

The pitch is that no single tech giant should own the standard. Innovation Labs says it will not hold the registration data itself. “There’s a lot of organ rejection to a hyperscaler releasing a standard and having that proprietary data,” interim boss Allie Kline told TechCrunch. The group says it is already trialling the system with several unnamed cloud giants.

An agent-shaped internet

The stakes are rising because agents are spreading fast, from Amazon’s revamped Alexa to enterprise tools, and they are already causing trouble. Researchers have tricked them into leaking private code and even running a full ransomware attack. Regulators are scrambling too, from China’s new agent rules to Delaware’s plan to give agents a legal identity.

Cerf is not sure the agent-run internet is inevitable. But he thinks people will try to build it anyway. “We are fundamentally lazy creatures,” he said. If an agent can do a job for us, we will let it.

For the past few years, our nation has been flooded with headlines declaring the demise of the college degree. This trend was exacerbated by COVID-19, which accelerated a decline in college interest.

I understand, really, I do. Tuition costs are rising. Student debt is real.

Artificial intelligence (AI) is also reshaping white-collar work by automating routine cognitive tasks, changing hiring patterns and increasing the use of AI tools in professional occupations. A 2025 Gallup survey found that AI use at work among U.S. employees nearly doubled from 21% in 2023 to 40% in 2025.

This is drawing many to a simple conclusion: a four-year college degree is no longer worth the time or money.

But the data, and the broader reality of how careers and life actually unfold, tell a different story.

Yes, the labor market for recent graduates has become more competitive. Yet college graduates still consistently outperform non-graduates in employment, earnings and long-term career resilience, according to new national data from the College Board Education Pays 2026 report.

But more importantly, a degree from a competitive college with a high graduation rate cultivates the ultimate asset in a rapidly changing economy: the ability to think critically. This includes being able to understand AI, as those who do will be better positioned to shape how it’s used ethically and responsibly.

That matters now more than ever.

Recent analysis from the Federal Reserve and labor economists shows that while the wage gap between graduates and non-graduates has narrowed, college graduates still maintain lower unemployment rates overall and stronger long-term job stability. A 2025 analysis from the St. Louis Fed found that from 2000 to 2025, workers with only a high school diploma consistently faced unemployment rates at least 2.3 percentage points higher than workers with bachelor’s degrees.

Even amid a softer hiring market, the advantage remains clear. Data cited by Goldman Sachs and other labor researchers showed unemployment for young non-college workers hovering around 7% in 2025, compared with roughly 4.6% for recent college graduates.

That is not a meaningless difference. In a large economy, a few percentage points represent millions of jobs.

Critics often focus narrowly on whether college guarantees a job immediately after graduation. That framing misrepresents the real purpose of higher education. College is not merely vocational training. It is preparation for a lifetime of economic and intellectual change.

The modern workforce is evolving too quickly for any technical skill to remain permanent. Entire industries now transform within a decade. Many students entering college today will eventually work in jobs that do not yet exist. In this environment, being able to think critically becomes the ultimate career skill.

A strong college education teaches students how to analyze information, communicate clearly, solve unfamiliar problems, conduct research, collaborate with different kinds of people, and learn independently. Those capacities transfer across different industries and technologies.

Ironically, the rise of AI may make these human skills even more important. Employers increasingly value workers who can think critically, interpret nuance and make judgments machines cannot easily replicate, according to Western Governors University, which surveyed more than 3,000 employers. Technical skills may evolve every few years; the ability to learn and think critically endures. According to McKinsey, “Human skills will matter more in the age of AI.”

The ability to think and process information is also why college graduates tend to weather recessions better over the course of their careers. Historically, workers with higher educational attainment have experienced lower unemployment during recessions and often recover faster in labor market recoveries, though this advantage varies by industry, age, and economic cycle. In 2024, unemployment for bachelor’s degree holders was 2.5%, compared with 4.3% for high school graduates and 6.1% for people without a diploma, according to the U.S. Bureau of Labor Statistics.

Of course, higher education must confront legitimate concerns about affordability and workforce alignment. There’s nothing wrong with questioning college directly after high school if a student is interested in pursuing a low-demand degree with high debt or if the student has yet to define a clear career goal.

But seeing college as only a trade school, in my opinion, is the wrong way to look at it. There are tremendous educations available where financial aid is available to help those who need it to meet the demands of higher education costs. There are wonderful State Schools and City Schools that are great choices for students.

This is an endorsement of a 4-year college degree, at a competitive school, to learn how to think critically, for a lifelong ability to learn new things. One thing we do know about the future is that we will need a population that has the ability to synthesize information quickly and accurately.

The real question is not whether college guarantees success. Nothing does.

The question is whether developing analytical ability, communication skills, flexibility, and intellectual independence still matters in an uncertain economy.

I am here to say they do. Perhaps more than ever.

The future will belong not simply to people who know things, but to people who can keep learning new things. College, at its best, remains one of the strongest environments for building that habit.

A college degree and a stable career can benefit generations. Earning a college degree is linked to longer, healthier lives, higher incomes, greater civic participation and better career alignment. While economic benefits are substantial, the lifestyle advantages extend to health, social engagement and personal fulfillment.

And that is why it is still worth it.

• Zoom patches critical improper input validation flaw in multiple Windows clients and SDKs that allowed remote account takeover
• Additional high‑severity bugs fixed include CVE‑2026‑53410 (TOCTOU race condition), CVE‑2026‑53409 (privilege management flaw), and CVE‑2026‑53411 (input validation issue)
• All vulnerabilities were found internally, with no evidence of exploitation; users are urged to update Zoom Workplace and related products to the latest versions

Zoom has patched a critical-level vulnerability in multiple products that allowed threat actors to take over people’s accounts remotely.

In a security advisory, Zoom said it fixed an Improper Input Validation bug plaguing Zoom Desktop Client for Windows (before version 7.0.0), Zoom VDI Client for Windows (before versions 7.0.10, 6.6.15, and 6.5.18), and Zoom Meeting SDK for Windows (before version 7.0.0). It did not go into more details on how the flaw works.

Through the looking glass: Modern car technologies such as license plate readers and self-driving telemetry systems have been raising privacy concerns for some time now. However, a San Diego company’s fleet observation devices might bring car surveillance to a new, uncomfortable level, especially when encountered by unsuspecting renters.

A woman recently took to social media after discovering that her Audi rental car’s dashboard contained a camera recording her every move. It also gave verbal reminders to wear a seat belt and avoid unsafe driving behavior.

The end of the video reveals that the camera was built by Lytx, which supplies its surveillance system to rental fleets. The company’s YouTube videos advertise the cameras as a safeguard that effectively minimizes distracted driving and other risky behavior.

– Wall Street Apes (@WallStreetApes) July 14, 2026

According to Lytx’s website, the cameras record events both in front of and behind the wheel using an AI-powered system that learns to detect when drivers use phones, smoke, eat, drink, follow other vehicles too closely, or ride without seat belts. They can record up to 400 hours of live video feeds and deliver verbal warnings to discourage unsafe driving. Fleet operators can set the cameras to record continuously or only in response to certain triggers.

However, the woman who posted the warning about the camera, calling it “the eye of Sauron,” was unaware that the rental car featured one before stepping into it. Lytx’s website shows its cameras almost exclusively in cargo trucks, and Audi is not listed among the company’s partners. The dealer that lent out the Audi might have installed the camera without the car manufacturer’s involvement.

While Lytx’s cameras appear to be legal in most US states and drivers are assumed to have consented by driving equipped vehicles, their legal standing regarding privacy remains unclear. Most people probably have a reasonable expectation of privacy while driving, even in a rental car.

Furthermore, the woman in the video noted that she is a medical professional who conducts private conversations with patients while driving. Recording those conversations might constitute a HIPAA violation.

Prior incidents of cars observing their drivers have mostly involved electric vehicles and self-driving cars such as Teslas and Waymos. In 2023, Tesla employees were found to be sharing deeply private and intimate videos captured on owners’ vehicles, some of them taken inside garages. More recently, a Waymo car stopped and contacted the police when its occupants fired toy guns from the vehicle. Systems such as Lytx’s raise the possibility that similar surveillance measures might be installed in almost any vehicle.

As you walk into XPeng’s Munich showcase event, you’re greeted by, I kid you not, a giant wooden Trojan horse. Not exactly a subtle message from a Chinese brand announcing its first-ever global release of an electric vehicle, right in the backyard of the German auto industry.

It’s hard to believe that XPeng was founded just shy of 12 years ago. Yet by 2020 it was already shipping EVs to Norway, marking the start of the Chinese company’s European journey. Today, alongside cars, it has robots and flying cars in its commercial product portfolio.

Look at the top 10 EV manufacturers in China by volume, and you won’t find XPeng, but it’s growing and has forged a bigger reputation outside of its home country. Now it wants to go global with its latest model, the L03, the brand’s first new car that will launch in 60 countries across Europe, Latin America, the Middle East, and the Asia-Pacific.

The L03 is a big play for XPeng because this is its “budget” model, starting at €35,600 (about $40,000), priced to sit below its G6 Tesla Model Y competitor, and to sell in volume.

The base-level XPeng L03 in Munich, complete with Trojan horse in the background.

Jeremy White

Yes, the L03 is the company’s mass-market play. Despite the keen pricing, XPeng has sought to make the specs attractive: a claimed WLTP 320-mile range; fast charging from 10 to 80 percent in 20 minutes; panoramic glass roof; heated and cooled massage seats; 256-color ambient lighting; brushed metal speaker covers; an impressive 0.228 drag coefficient to squeeze out more range; smart parking; a 15.6-inch 2.5K central screen; 27-inch HUD; AI-powered voice control; and even Google Maps built in.

All this and more come as standard, whether you go for the vanilla model, the Long Range, AWD, or Ultra. The phrase XPeng keeps using for this embarrassment of riches is “beyond class.” It wants the L03 to go toe-to-toe with EVs in the segment above it—cars like the Volkswagen ID.4.

Performance? Well, the five-seat, 4,650-mm L03 can hit 0 to 60 mph in just 4.5 seconds on the top models, but this drops to 7.5 seconds on the Standard Range base version.

Photograph: Courtesy of XPeng

The two technologies work in tandem.

Can you run Android Auto without Bluetooth?

Why does wireless Android Auto need Wi-Fi?

Pexels: Tim & Martin Klement

What happens if your car doesn’t have wireless Android Auto?

What are the disadvantages of wireless Android Auto?

DimasGermany/Shutterstock