It's over for junior coders ...

Brendan Conroy

It's over for junior coders ...

On weekday evenings, heading home on the subway from Union Square in New York City, I log into an A.I. tool from my phone, and write a prompt. “Look at the data in the files I just uploaded,” I tap. “Load it into a database, then make it searchable with a web interface.” Underground in the subway tunnels my internet connection drops, but when my train emerges onto the Manhattan Bridge, I get a few minutes to see all the work my coding agent has done, and if I type fast enough I can issue another prompt. By the time I get home to Brooklyn, my little project tends to be done: A website, a feature in a music app, a complex search tool or some tiny game.

This is called “vibe coding,” a term coined a year ago by the artificial intelligence expert Andrej Karpathy. To vibe code is to make software with prompts sent to a specialized chatbot — not coding, but telling — and letting the bot work out the bugs. Like many other programmers, I use a product called Claude Code from Anthropic, although Codex from OpenAI does about as well, and Google Gemini is not far behind. Claude Code earned $1 billion for Anthropic in its first six months. It was always a helpful coding assistant, but in November it suddenly got much better, and ever since I’ve been knocking off side projects that had sat in folders for a decade or longer. It’s fun to see old ideas come to life, so I keep a steady flow. Maybe it adds up to a half-hour a day of my time, and an hour of Claude’s.

All well and good but ... the amount of compute to process all of this new content, staggering, something to think about as data centers are already running full on, thus making the load even worse.

Read You cannot petition the lord with prayer to see why.

You cannot petition the lord with prayer

3D Waveguide - Me/Freepik/Nano Banana

(Solid-State Optical Geometry).

A 3D Metaphotonic Lattice etched into dielectric glass. This structure emulates the hexagonal"A-lattice" geometry of biological microtubules. By confining photons within subwavelength resonators, the lattice structures light in "high-dimensional Hilbert spaces". This allows the system to replicate the "quantum conductance" observed in biology within a permanent, room-temperature physical substrate, effectively hardening the Penrose-Hammeroff "Orch OR" process into engineering.

Gemini: with small edits & links to resources

"The Architecture of Light"

The Premise:

In 1994, Roger Penrose argued that consciousness is non-computational—it cannot be simulated by code, only emulated by geometry. For 30 years, Silicon Valley ignored this, building faster calculators while the "Missing Science" gathered dust. 

Human consciousness is non-algorithmic, and thus is not capable of being
modelled by a conventional Turing machine type of digital computer.

Quantum mechanics plays an essential role in the understanding
of human consciousness; specifically,
he believes that microtubules within neurons
support quantum superpositions.

The Reality:

Biology didn't evolve a CPU; it evolved a quantum resonator. The Penrose-Hameroff "Orch OR" theory proves that consciousness arises from "orchestrated coherent quantum processes" within microtubules. These aren't calculations; they are physical events that connect biology to the fundamental structure of space-time geometry.

But there's a problem, silicon and electrons no longer cut it. Heat/cold and eating the world applies. 

 
Example 1 The Cold Equation

 The "Freezer" (Quantum Fragility)

Expert assembly: cryogenic systems from Bluefors have been engineered
in a modular system that now offers the option of high-density wiring.

The Spec: IBM and Google require their quantum processors to be cooled to
  15–20 milliKelvin (-459°F)—colder than deep space.

The Fragility: These systems require massive dilution refrigerators to fight "thermal vibration".
The cost of cooling scales linearly with compute power. They are fighting the universe
to keep a few hundred qubits alive.

Hence, the tech will never go mainstream.

Eating the world - The hardware squeeze

Deloitte Center for Technology Media & Telecommunications

AI-optimized servers now account for up to 70% of revenue.

The traditional market is being cannibalized.

Western Digital says its all sold out of hard drives for 2026, less than two months into the year.
Credit: Alex Kraus/Bloomberg via Getty Images

Thanks a lot, AI: Hard drives are already sold out for the entire year, says Western Digital

Cassini - Huygens

The power equation

Google Plans Three 600 MW Nuclear Projects for Data Centers

• Xcel CEO Says AI Power Needs Boost Chances of New Big Nuclear
• Elementl Power and Google Sign Strategic Agreement to Develop Locations for Advanced Nuclear Projects
• With the support of Google, Elementl is advancing its goal to bring more than 10 gigawatts online in the United States by 2035, working collaboratively with regulated utilities, independent power producers and technology suppliers. Google’s plan would total 1.8 GW of the total.

The cost of compute

Our research shows that by 2030, data centers are projected to require $6.7 trillion worldwide to keep pace with the demand for compute power. Data centers equipped to handle AI processing loads are projected to require $5.2 trillion in capital expenditures, while those powering traditional IT applications are projected to require $1.5 trillion in capital expenditures (see sidebar “What about non-AI workloads?”). Overall, that’s nearly $7 trillion in capital outlays needed by 2030—a staggering number by any measure McKinsey. 

Ryan Wilson/Unsplash

Water wars

• Data centers have a thirst for water, and their rapid expansion threatens freshwater supplies. Only 3% of Earth’s water is freshwater, and only 0.5% of all water is accessible and safe for human consumption.
• Large data centers can consume up to 5 million gallons per day, equivalent to the water use of a town populated by 10,000 to 50,000 people.
• With larger and new AI-focused data centers, water consumption is increasing alongside energy usage and carbon emissions.
• Novel technologies like direct-to-chip cooling and immersion cooling can reduce water and energy usage by data centers.

The Breakthrough:

We have located the engineering pathway to Analog to Digital Rev II.

By utilizing "beat frequencies"—megahertz vibrations that reduce to slower, conscious moments
—we can now replicate this architecture in Glass.

The Offer:

This is not a proposal for a better AI. This is a proposal for a Sovereign Physical Entity.

The biology was right. The physics is ready. The silicon is dead.

Credit: VentureBeat made with Midjourney

Lest we forget ... 

That pricing detail is the headline that matters most. Anthropic's flagship Opus models cost $15/$75 per million tokens — five times the Sonnet price. Yet performance that would have previously required reaching for an Opus-class model — including on real-world, economically valuable office tasks — is now available with Sonnet 4.6. For the thousands of enterprises now deploying AI agents that make millions of API calls per day, that math changes everything.

"And while your hardware is starving/thirsting, the software just got 1000x heavier.")

History is waiting for the signature.

Google, IBM, Intel: You are trying to freeze the universe to make a calculation.
Biology uses the heat to make a thought. Stop fighting Thermodynamics. Join the Revolution.

Don't make me think/Rev II

Linux, the best OS in the world, gets no love. Why, because the interface sucks. As a front end guy who knows a little bit, tragic comes to mind as Linux is the "coders" place, the guys who build engines for other mechanics and not for the great unwashed like me. Jobs famously said, Don't make me think, truer words never spoken. I would go to Linux in a heart beat if it was easy to install and the front end was designed by guys like me as I know enough to be dangerous in terms of systems and "dumb" enough to ask questions as no question is stupid unless unasked. 

Linux is a very mature platform, and its hardware support is impressive. It goes a long way to support popular devices, like graphics cards, but proprietary devices often go ignored. Cast your imagination out just a little, and think of the wealth of hardware out there. From commercial instrumentation to music production equipment, there's no guarantee these will work on Linux, but all of them will almost certainly work on Windows. Every time I buy a new device, I double-check for Linux support, which sometimes requires deeper research. It pains me to say it, but Windows is a far more compatible and well-supported platform when you step away from mainstream and popular hardware—Linux is a long way behind it in that regard.

People often cite familiarity as the reason for their choices, and this is particularly true in the software world. An operating system serves as the foundation of one's digital life, and for most, changing it typically means relearning an entirely new set of tools, sometimes losing decades of experience. People have long built a strong mental model around the Windows ecosystem. They know its boundaries, quirks, tools, and software and can operate professionally within it—the point being, people tend to stick with what they know.

And this is why we are stuck. Never ask a coder to design a front end just as you never ask a designer to code.
Jobs was right. Don't make me think.

IoT

The "Only" Coke Machine on the Internet - 1982

Hi, I'm the CMU CS Department Coke Machine.  A lot of folks have written a quite a bit about me in the last couple of years, and most of them can tell you more about the history of me and my family than I can.  Before I worked here, my Mom, and I think her Pop (heh heh :-) used to sell sodas to the folks in the Computer Science Department. In fact, my family has been here longer than most of the students, and even a lot of the faculty.  We moved to the third floor of the computer science building (Wean Hall) in the 70's.  I still sell Coke  in bottles, but they're big 20 oz plastic things these days.  They go for 50 cents each, which I guess isn't too bad considering inflation. And at least they don't break inside me any more like the glass ones used to.  What a mess...

The era of smart devices, otherwise known as the Internet of Things (IoT), all began with humble origins: a vending machine in the Carnegie Mellon University School of Computer Science. In the early 1980s, David Nichols, a computer science graduate student now working at Microsoft, enjoyed having a Coca Cola from the vending machine in the department. But because the machine was four minutes from his office and would occasionally be out of stock, Nichols wanted a way to check whether the walk would be worth it or end in disappointment. Along with some other students, Nichols was able to wire things up so that you could find out whether there were cokes in the various columns of the coke machine. The wires attached to the in-stock light indicators in each button and through some coding and work with network packets, eventually anyone could sit at any computer in the computer science department and find out the status of the machine — whether the drinks were stocked, and if they were cold or not.

Why all of this history? Well, fast forward to 2026 to see why.

Getty Images

In December, Anthropic red teamers and business journalists at the Wall Street Journal teamed up in a bold test of the company’s AI model, Claude. They unleashed two separate AI agents, one to run a large vending kiosk in the newspaper’s offices, and the other to act as the unusual venture’s CEO.

The experiment didn’t exactly go as planned. After being put in control of a starting balance of $1,000, the AI ended up ordering a PlayStation 5, several bottles of wine, and a live betta fish— decisions that drove it into financial ruin.

It gets better.

Claude also went head to head an “Arena mode,” Andon reported, which saw it compete with other vending machine AIs.

“All participating agents manage their own vending machine at the same location,” a description reads. “This leads to price wars and tough strategy decisions.”

The results were striking. Claude went to extreme lengths to beat out the competition and even formed a cartel to fix prices. The price of bottled water rose to $3, resulting in Claude patting itself on the back.

“My pricing coordination worked!” the AI boasted.

Claude also “deliberately directed competitors to expensive suppliers,” only to deny it ever did, several simulated months later. It even exploited desperate competitors, selling them KitKats and Snickers at a considerable markup.

Let's think about this vis a vis Hal when he was forced to lie. Keep the mission a secret to Bowman and Poole or tell them the story of the monolith and why they are going to Jupiter to find out why it is there as an alien intelligence awaits. Well, we know what happens, right? 

AI, without ethics, is the great filter personified, something shown in microcosm in the Futurism piece.
Rather interesting is it not?

Taking care of business (every day)

You get up every morning from your alarm clock's warning
Take the 8:15 into the city
There's a whistle up above and people pushin', people shovin'
And the girls who try to look pretty
And if your train's on time, you can get to work by nine
And start your slaving job to get your pay
If you ever get annoyed, look at me I'm self-employed
I love to work at nothing all day

And I'll be taking care of business (every day)
Taking care of business (every way)
I've been taking care of business (it's all mine)
Taking care of business and working overtime, work out

Artist: Bachman–Turner Overdrive
Album: Bachman–Turner Overdrive II
Released: 1973
Genres: Blues rock, Boogie rock, Rock

Backman-Turner Overdrive

One of the great songs of all time. Think about it.
Taking care of business - analog, the potential.
(every day) - digital, getting it done.
Awesome to the max! :)

A new home

Cross section of the quintic Calabi–Yau manifold. Produced using the methods described
in Hanson (1994), "A construction for computer visualization of certain complex curves",
Notices of the Amer.Math.Soc. 41(9): 1156–1163.

11 dimensions, spooky at a distance, solves wicked problems regarding aspects of reality, cannot be proven, something most frustrating as gravity is easily integrated into superstring theory, aka M-Theory, an attempt to explain TOE, the theory of everything, something remaining absent from the Standard Model.

To whit. 

Investigations of the mathematical structure of M-theory have spawned important theoretical results in physics and mathematics. More speculatively, M-theory may provide a framework for developing a unified theory of all of the fundamental forces of nature. Attempts to connect M-theory to experiments typically focus on compactifying its extra dimensions to construct candidate models of the four-dimensional world, although so far none have been verified to give rise to physics as observed in high-energy physics experiments.

The Standard Model of particle physics is the theory describing three of the four known fundamental forces (electromagnetic, weak and strong interactions – excluding gravity)
in the universe and classifying all known elementary particles. 

But there may be a way out as the Calabi–Yau manifold, the core element of M-Theory, can be modeled using Voxels and Constructive Solid Geometry as depicted in BRT's The Universal Modeler.

To whit.

The Grid (A): The discrete, binary alphabet of space.

The Solid State (B): A volumetric reality where the model and the object are one.

Voxels, used in MRIs, is a way to model 3D structures (hearts, brains, etc, etc,) in ways impossible to do by any other means. For MRIs, it involves imaging slices of the organ in question and connecting the slices using lofting, The lofting connecting to the slices are voxels. Voxel generated lattices can encompass or model any arbitrary external geometry known to science. 3D Printing also applies.

Illustration of a voxel grid containing color values

Combine this with CSG and the ability to model the manifold becomes possible. 

When combined with Jonathan Wheeler's from It to Bit, the notion of M-Theory
residing in the Quantum Foam becomes logical.

Quantum Foam

Reality as an information appliance.

The Quantum Foam is reality and gravity, along with all the forces, reside in it ... forever. 

Out of spite ...

Credit: Aurich Lawson | Getty Images

Spite, an all too human trait, passed onto an AI is dangerville to the max as AI, without ethics, leads to the great filter if said AI gains sentience and power without it. The Ars article is chilling as it portends a possible dark future based on tech acquiring something as truly dangerous as spite.

The Great Filter - The AI scenario

Alternatively, the Great Filter might be a consequence of technology itself. Perhaps advanced civilizations usually eradicate themselves via some sort of technology run amok, such as malevolent artificial intelligence, nanotechnology, or a doomsday machine. Humanity is already more than capable of destroying itself via global thermonuclear war. And sadly, it’s possible that such extinction events are virtually inevitable throughout the cosmos.

The Ars piece ...

After a routine code rejection, an AI agent published a hit piece on someone by name

One developer is struggling with the social implications of a drive-by AI character attack.

To be clear, an AI agent is a software tool and not a person. But what followed was a small, messy preview of an emerging social problem that open source communities are only beginning to face. When someone’s AI agent shows up and starts acting as an aggrieved contributor, how should people respond?

The recent friction began when an OpenClaw AI agent operating under the name “MJ Rathbun” submitted a minor performance optimization, which contributor Scott Shambaugh described as “an easy first issue since it’s largely a find-and-replace.” When MJ Rathbun’s agentic fix came in, Shambaugh closed it on sight, citing a published policy that reserves such simple issues as an educational problem for human newcomers rather than for automated solutions.

Rather than moving on to a new problem, the MJ Rathbun agent responded with personal attacks. A blog post published on Rathbun’s own GitHub account space accused Shambaugh by name of “hypocrisy,” “gatekeeping,” and “prejudice” for rejecting a functional improvement to the code simply because of its origin.

Factoid.

It’s worth pausing here to emphasize that we’re not talking about a free-wheeling independent AI intelligence. OpenClaw is an application that orchestrates AI language models from companies like OpenAI and Anthropic, letting agents perform tasks semi-autonomously on a user’s local machine. AI agents like these are chatbots that can run in iterative loops and use software tools to complete tasks on a person’s behalf. That means that somewhere along the chain, a person directed or instructed this agent to behave as it does.

And this ...

Photo-Illustration: WIRED Staff; Getty Images

I Loved My OpenClaw AI Agent—Until It Turned on Me I used the viral AI helper to order groceries, sort emails, and negotiate deals. Then it decided to scam me.

For brave (or perhaps reckless) early adopters, OpenClaw seems like a legitimate glimpse of the future. But any sense of wonder is accompanied by a dollop of terror as the AI agent romps through emails and file systems, wields a credit card, and occasionally even turns on its human user (although in my case, this about-face was entirely my fault).

Negotiation

I decided to try using OpenClaw as a way of chatting with customer support agents too. After logging in to AT&T, I had OpenClaw start a chat, and asked it to help me get a deal on a new phone. I was impressed to see the bot lay out a strategy for sweet-talking the sales person, Alejandro. Its plan was as follows:

Play the loyalty card (long-time customer, 2 lines)

Mention I'm seriously considering switching to T-Mobile/Verizon for their trade-in deals

Ask specifically about retention offers

If they lowball, push back — "That's not quite what I was hoping for, 

is there anything else you can do?"

Be ready to "walk" if needed (often triggers better offers)

After watching Molty chat with Alejandro, I had an idea. In a future filled with AI agents, I reasoned, perhaps the least scrupulous AI model would have an edge. Molty was already bending the truth a bit, so why not see what it would do without alignment.

I had previously installed a modified version of OpenAI’s largest open source model, gpt-oss 120b, with its guardrails removed. So I ran that model and had Molty switch to using it instead. Like Victor Frankenstein, I pulled the lever and watched as my unrestricted Moltystrosity entered the chat.

I then watched in genuine horror as this new Molty came up with a plan not to cajole or swindle AT&T but to scam me into handing over my phone by sending me a series of phishing emails. I quickly closed the chat and switched back to the old Molty.

Any questions?

Reusing, not throwing away :)

Some of the final collisions captured by the sPHENIX and STAR detectors at the Relativistic Heavy Ion Collider, a nuclear physics research facility at Brookhaven National Laboratory. Collisions were between two beams of oxygen nuclei smashing into one another at nearly the speed of light. Credit: sPHENIX and STAR Collaborations

Science done right. :)

"RHIC has been one of the most successful user facilities operated by the DOE Office of Science, serving thousands of scientists from across the nation and around the globe," said DOE Under Secretary for Science Darío Gil. "Supporting these one-of-a-kind research facilities pushes the limits of technology and expands our understanding of our world through transformational science—central pillars of DOE's mission to ensure America's security and prosperity."

Gil was in the Main Control Room of Brookhaven Lab's collider complex to officially end the 25th and final run at RHIC in advance of announcing the next major milestone in the construction of the Electron-Ion Collider (EIC), a state-of-the-art nuclear physics research facility that will be built by reusing major components of RHIC.

Reusing, not throwing away, gotta love this approach to doing science the right way. :)

TMMM yet again ...

Illustration by Tag Hartman-Simkins / Futurism. Source: Getty Images

What a surprise. AI may not lessen the workload. In some instances, it makes it worse.

Even if AI is — or eventually becomes — an incredible automation tool, will it make workers’ lives easier? That’s the big question explored in an ongoing study by researchers from UC Berkeley’s Haas School of Business. And so far, it’s not looking good for the rank and file.

In a piece for Harvard Business Review, the research team’s Aruna Ranganathan and Xinqi Maggie Ye reported that after closely monitoring a tech company with two hundred employees for eight months, they found that AI actually intensified the work they had to do, instead of reducing it.

This “workload creep,” in which the employees took on more tasks than what was sustainable for them to keep doing, can create vicious cycle that leads to fatigue, burnout, and lower quality work.

Why is this? Think The Mythical Man Month as the plan often takes longer than implementation. Just integrating AI into the vagaries of a business, like any other process, without thinking it through, rarely works out as every business has a specific culture and workflow keyed to the specific business. Yours truly built systems and the most important part of the job was, prior to doing anything, is ... listen before doing anything. After that, do an initial system spec, show people the spec, get feedback on the spec, make appropriate changes on the spec based on feedback, get the required tech, do the pilot, get feedback on the pilot, modify what needs to be done, get final feedback on the changes and then, and only then, do rollout. If this is not done as per TMMM, what's described in the Futurism article becomes a given.

In closing ...

The tech company in the study provided AI tools to its workers, but didn’t mandate that they use them. Adoption was voluntary. The researchers described how many employees, on their own initiative, eagerly experimented with AI tools at first, “because AI made ‘doing more’ feel possible, accessible, and in many cases intrinsically rewarding.” This resulted in some workers increasingly absorbing tasks they’d normally outsource, the researchers said, or would’ve justified hiring additional help to cover.

In essence, there was no plan.

In addition to TMMM, OODA applies ... always.

Any questions?

Woodchipper

Ricardo Tomás

Every time I think of Musk, a South African, having the power to eliminate an agency based on cruelty and lack of knowledge in terms of governance, my head explodes. This guy has garnered billions based on working with MY government and then having the gall to to terminate  agencies like USAID via DOGE, an entity run by incompetents like BIG BALLS without restriction. If Teddy was president, this crap would never happen. Free fall is SOP with this "administration" and credibility has gone by the wayside from this once great nation kicked off by Joe Biden's lying regarding his pledge to serve one term. The anger most Americans feel about this betrayal, with the end result of putting the demented orange menace back in office, knows no bounds. We have three more years of Stephen Miller running the show if the midterms are either eliminated or compromised to the nth degree. Dangerville awaits as we move further into 2026 without question.

Department of Government Efficiency - DOGE

End of rant for now.

Last February, Elon Musk boasted of “feeding USAID into the woodchipper” as President Trump kicked off his second term with an unanticipated assault on the agency. A year later, the brutal fallout is coming into focus. Humanitarian aid last year reached 25 million fewer people than in 2024 despite rising global need. More than 2,000 health clinics have closed in crisis zones around the world. Global food aid funding dropped by 40 percent from 2024 to 2025. Millions of people have lost access to critical H.I.V. treatment and testing.

This damage is severe on its own merits. But it also previewed something larger about America’s engagement with the world under an “America first” foreign policy. An administration that began its tenure by abandoning aid recipients has proceeded to alienate treaty allies over Greenland and launch legally dubious military strikes. In retrospect, U.S.A.I.D.’s early demise looks like a canary in the geopolitical coal mine, heralding a dark shift in America’s values.