So Much Tech News: The Week of 10/18/2026

Alex Michael

Oct 25, 2026

Share

This Week in the News…

Scientists who discovered a third form of magnetism win major physics prize (Knowledge Required: 🧠🧠🧠🧠)

Disney agrees to hire the majority of Character.AI’s workforce 🧠

Anthropic says it’s committing 5% of its shares to a ‘public benefit fund’ 🧠🧠

How far can we push computing? 🧠🧠🧠🧠🧠

Musk pledges to give ‘every kid in America’ one of his SpaceX shares 🧠

People are using AI to make their smart TVs dumb again 🧠🧠

AI Proved the Quasi-Riemann Hypothesis. What Happens Now? 🧠🧠🧠🧠🧠

What to Know About the Obscure Trust Looming Over Anthropic’s IPO 🧠🧠🧠

Diffusion at Scale: Introducing Mercury 3 🧠🧠🧠🧠

US banning Chinese AI DeepSeek from government agency IT systems 🧠🧠

How abliterated models can get you pwned 🧠🧠🧠🧠

TikTok staff raised child safety concerns, but executives reportedly failed to act 🧠

eMudhra Intermediate CA Compromise 🧠🧠🧠🧠

How I got over 300 Tor exit nodes removed 🧠🧠🧠🧠

Plants can speak using electricity, so this startup built an AI model to decipher their language 🧠🧠🧠

Facebook now limits you to two links a month. Why are you still there? 🧠

Google has joined Amazon in attacking Nvidia’s stronghold with AI app developers 🧠🧠🧠🧠

SpaceX Q3 2026 Earnings Call Transcript 🧠🧠🧠

Eight Bits of Trust; eMudhra’s CA keys cracked in milliseconds (Follow up) 🧠🧠🧠🧠

Feds Suspect Ransom Negotiator Ran ShinyHunters Extortion Site 🧠🧠

In Hilbert Space, All Things Are Quantumly Possible 🧠🧠🧠🧠🧠

Claude Duet — Get back to what needs you 🧠🧠

‘This Is Not a Drill’: Critical PHP Vulnerabilities Are Setting the Internet on Fire 🧠🧠🧠🧠

Mark Zuckerberg Agrees to Stay in California Despite Billionaires’ Tax Ballot Measure 🧠🧠🧠

Several Linux Kernel Vulnerabilities Accidentally Posted To Public Kernel Mailing List 🧠🧠🧠

eMudhra to Exit global SSL/TLS Certificate Business; Will Sharpen Focus on Core Enterprise Solutions and trust services other than SSL/TLS (Follow Up) 🧠🧠🧠

Mini brain organoids ‘reading’ Braille marks key step towards biohybrid computing 🧠🧠🧠🧠

OpenAI working on new $30B raise at $1.25T valuation ahead of IPO 🧠🧠

The Social Society Helping Women Join the AI Boom 🧠

Warhound raises $200M to engineer war dogs and arm them with Anduril gear 🧠🧠

Scientists discover a hidden messaging system deep inside the brain 🧠🧠🧠🧠

A tiny defect in a crystal reveals a big quantum surprise 🧠🧠🧠🧠🧠

Updated Google privacy policy says your Gmail might be used for advertising 🧠🧠

Securing our frontier model training environment 🧠🧠🧠🧠🧠

Block’s new voice model sounds eerily human 🧠🧠

Ukraine’s Military Starlink Is Working Again, Kyiv Says 🧠

Why One of the World’s Greatest Mathematicians Is on a ‘War Footing’ 🧠🧠🧠

iPhone — Today, we hacked what we’d set out to hack: the iPhone 🧠🧠🧠

I DON’T UNDERSTAND QUANTUM PHYSICS 🧠🧠🧠

1. Scientists who discovered a third form of magnetism win major physics prize (ScienceDaily) 🧠🧠🧠🧠

By Libor Šmejkal - Provided by the author, CC BY-SA 4.0, https://commons.wikimedia.org/w/index.php?curid=145420926

Three physicists, Jairo Sinova, Libor Šmejkal, and Tomas Jungwirth, have received the 2026 Europhysics Prize for establishing altermagnetism, a third fundamental form of collinear magnetic order alongside ferromagnetism and antiferromagnetism.

Unlike ferromagnets, where atomic magnetic moments align to produce an overall magnetic field, altermagnets have opposing magnetic moments that cancel out, similar to antiferromagnets. However, their unique crystal symmetries produce electronic properties typically associated with ferromagnets, including spin-polarized electronic states.

This combination could enable faster, more energy-efficient electronics, particularly in spintronics, which uses electron spin rather than just electrical charge to process and store information.

The researchers developed a comprehensive theoretical classification in 2022, identifying hundreds of potential altermagnetic materials. Subsequent experiments confirmed the phenomenon in several materials, establishing an entirely new area of condensed matter physics with potential implications for quantum computing, superconductivity, and advanced electronics.

A legal filing spotted by Bloomberg revealed more details about the agreement between Character.AI and Disney, with Disney agreeing to make employment offers to at least 60 Character.AI employees — the majority of the workforce — paying them an estimated $60 million in annual compensation.

Anthropic said that it will commit 5% of its shares to the fund — structured as a wholly-owned entity within the company — with the option to contribute an additional 2% for every 50% increase in Anthropic’s value. And it offered some broad-strokes ideas for how this money might actually be used, such as “large donations to traditional nonprofits or direct distributions to specific populations.”

Source: MIT Technology Review, “How far can we push computing?” (October 19, 2026).

The computing industry is approaching the limits of conventional semiconductor scaling, but we remain extraordinarily far from the fundamental limits imposed by physics. Historically, computing improvements came from shrinking transistors, making computers faster, smaller, cheaper, and more energy-efficient. Today, those improvements are slowing, and progress increasingly depends on building larger systems with more processors, memory, electricity, and cooling infrastructure.

The article argues that future breakthroughs may require fundamentally new computing architectures rather than continued refinement of existing silicon technology.

1. Why computing improvements are slowing

In 1959, physicist Richard Feynman delivered his famous lecture, There’s Plenty of Room at the Bottom, arguing that the laws of physics permitted manipulating matter at extremely small scales. Decades of semiconductor development subsequently demonstrated this principle.

Several major trends historically drove computing improvements, but nearly all are slowing:

Moore’s Law: Transistor density historically doubled approximately every two years, but improvements have become slower and more expensive.

Dennard Scaling: Shrinking transistors once reduced their power requirements while increasing performance. This effectively ended around 2004 as heat dissipation became a major limitation.

Koomey’s Law: Improvements in the number of computations performed per joule of energy are slowing.

Kryder’s Law: Improvements in magnetic storage density have slowed substantially.

Keck’s Law: Growth in the information capacity of conventional optical fibers is flattening.

As a result, performance improvements increasingly depend on scaling entire systems rather than individual components.

AI exemplifies this shift. Training compute has approximately doubled every six months since 2010, far exceeding improvements from semiconductor miniaturization. Meanwhile, six major hyperscalers are projected to spend more than $700 billion on computing infrastructure in 2026, roughly twice their 2025 expenditure.

2. How far are we from the physical limits?

Despite slowing technological progress, existing computers remain extraordinarily far from what physics theoretically permits.

Transistor density has increased approximately 10 billion-fold.

Computational speed has increased approximately one trillion-fold.

Energy consumption per elementary operation has fallen by roughly 17 orders of magnitude.

In 2000, physicist Seth Lloyd calculated the theoretical performance of an ultimate laptop: a hypothetical computer consisting of one kilogram of matter contained within one liter.

By exploiting its entire mass-energy for computation, such a computer could theoretically perform approximately 5 × 10⁵⁰ operations per second, 35 orders of magnitude faster than today’s leading chips.

However, accomplishing this would require operating at temperatures around one billion degrees, making such a system completely impractical with current technology.

The broader point is that fundamental physical limits remain vastly beyond our current engineering capabilities.

The article identifies four theoretical constraints that govern the ultimate capabilities of computing.

Quantum mechanics determines the maximum number of operations a physical system can perform using a given amount of energy.

Today’s chips operate approximately 35 orders of magnitude below the theoretical limit for a one-kilogram computer.

Closing this gap would require fundamentally different computing architectures capable of exploiting far more of matter’s physical properties.

In 1961, physicist Rolf Landauer demonstrated that erasing one bit of information must dissipate a minimum amount of energy, approximately 2.9 × 10⁻²¹ joules at room temperature.

Modern chips consume several hundred thousand times more energy during typical elementary switching operations.

One potential solution is reversible computing, which avoids logically irreversible information erasure and could dramatically reduce energy losses.

3. Information storage — Bekenstein bound

The Bekenstein bound, derived from black-hole thermodynamics, establishes the maximum amount of information that can exist within a bounded region of space, given its size and energy.

For Lloyd’s hypothetical laptop, this corresponds to approximately 10⁴² bits, nearly 30 orders of magnitude beyond what equivalent-sized modern storage devices can achieve.

DNA storage demonstrates how information could be stored at molecular scales, although writing and retrieving information remain too slow and expensive for mainstream computing.

The Holevo bound limits the amount of classical information that quantum states can communicate.

The article estimates that an ideal optical channel using one watt of power could theoretically transmit approximately 10¹⁷ bits per second.

Modern optical interconnects operate at approximately one-millionth of this rate for comparable power. However, advances in photonics and chip-integrated optical communications are narrowing the gap.

Of the four limits, communication efficiency may be the first that the industry approaches.

4. Technologies that could revolutionize computing

The article highlights several emerging approaches that could eventually overcome limitations in conventional computing:

Exploits quantum states to perform certain calculations, potentially enabling capabilities beyond classical architectures. However, today’s qubits are fragile and require extensive error correction.

Avoids unnecessary information erasure and recovers energy during switching. Vaire Computing tested an early chip incorporating energy-recovery techniques in 2025.

Encodes information at molecular scales, potentially achieving extraordinary storage densities. Its biggest obstacles are slow read/write speeds and high costs.

Use photons instead of electrical signals to transmit information, potentially improving bandwidth and energy efficiency. Companies including Nvidia and Broadcom are already shipping technologies that integrate optical communications close to computing chips.

Each approach faces substantial engineering challenges, and none guarantees commercially practical improvements in the near term.

5. The central argument: Build from the bottom up

The author distinguishes between two constraints on computing progress:

The practical limits of infrastructure expansion, including electricity, land, manufacturing capacity, cooling, and financial investment.

The fundamental physical limits of how efficiently matter and energy can compute, store, and transmit information.

Historically, computing advanced in both directions: manufacturers built larger systems while individual components became smaller and more capable.

Today, improvements increasingly come from building larger systems, bringing the industry closer to practi