A Plain-Language Edition of Quantum from the Ground Up

Einstein called it spooky action at a distance. He meant quantum entanglement, and
he thought it couldn't be real. He was wrong!

I published a plain-language edition of Quantum from the Ground Up today, built for that reader.

The original book is written for people with a technical background who haven't taken a graduate course in quantum mechanics. It works well for that reader. It does not work for someone who has never had a reason to think about qubits, cryogenics, or post-quantum cryptography and just wants to know what the technology actually does.

The two audiences want different things. Someone building a career in the field wants precision: exact fidelity numbers, fabrication steps, the difference between a physical qubit and a logical one. Someone reading a headline about a new quantum computing milestone wants to know what it means and whether it matters to them. The plain-language edition is built for the second person.

It covers the same 19 chapters as the technical edition, at a fifth of the length: 13 pages instead of 73. It leaves out the math, the fabrication chemistry, and the dense statistics tables. It keeps the facts that matter: what quantum computers actually do right now, why companies are spending billions of dollars building them, why your bank account and your medical records are already part of this story, and where the actual jobs are for people without a physics degree.

Both editions are free. Both get updated quarterly as new research and hardware announcements come out. The next technical edition publishes tomorrow.

You can find current and older versions of each in this folder.

If you have been putting off understanding quantum computing because every explanation you found assumed a physics background you don't have, this edition removes that excuse.

What Mrs. Anderson’s High School Chemistry Class Taught Me About Units

I found a bunch of posts I started writing years ago but never finished. Here’s one of them.... just finished.

I remember Mrs. Anderson at the blackboard in my high school chemistry class, writing out a conversion problem - something like converting a volume in liters to milliliters to moles. She worked it by stacking fractions, one after another, each one arranged so a unit in the numerator of one fraction matched the unit in the denominator of the next. Then she crossed them out in pairs until only the answer's units remained. "Learn a few equations," she said, "and you can solve just about anything." She was not talking about memorizing formulas. She was talking about what she called factor labeling.

Factor labeling - today more commonly called a more fancy dimensional analysis - treats units as algebraic objects. You multiply and divide them the same way you multiply and divide numbers. A conversion factor like 1,000 milliliters per liter is really just the number one, with some units, so multiplying by it changes the label without changing the quantity. Chain enough of these factors together and the units in between cancel, leaving you with exactly the unit you wanted.

The value of the method is not speed. It is error detection. In high school it helped in chemistry (not my favorite subject). In college physics (loved it) it became something I depended on, once problems started combining velocity, acceleration, force, and energy in the same calculation and a single wrong exponent could hide inside an otherwise reasonable looking number..

Here’s a simple example. In circuit analysis, say a resistor carries 25 milliamps (I) at 12 kilohms ® and you need to figure the power (P) dissipated in watts. Run the raw numbers straight through P = I²R without tracking units and you get 7,500 - off by a factor of a thousand from the real answer. Run it through factor labeling instead: convert 25 milliamps to 0.025 amps and 12 kilohms to 12,000 ohms before multiplying, and the units confirm the answer lands in watts: 7.5 watts.

I have used this same check in electrical engineering courses, in circuit analysis, and in grading student calculations for capstone projects. A wrong answer with clean unit cancellation is rare. A wrong setup almost always leaves a stray unit sitting where it should not be.

Fifty plus years after high school, I always check my units before I trust my numbers. Thanks, Mrs. Anderson!

Forty Years, No Matching Belt

Plotting World Fiber Domination in a tie at Microsoft 1998

A photo from Microsoft in 1998 - me in a tie, with a guy whose name I do not remember, 
sketching out fiber optic network technology plans at a whiteboard. This was my first time in Redmond and I wore that tie once, on my first day there, before I noticed nobody else there bothered with one. Back at the college, dress shoes were still a thing at least for me, and I never once got the belt and shoes to match. I did not even know that was a fashion rule. One would be brown, the other black….

Engineers a generation before mine wore bow ties and short sleeve shirts for a practical reason: a long tie or a loose cuff could catch in a running machine on the shop floor. That risk faded as engineering work moved from the floor to the desk, and clothing followed the same path from function to convention. By the early 1990s, tech had already dropped the tie. Engineering classrooms took another half decade or so to catch up.

The tie disappeared from teaching in the 1990s, replaced by a collared shirt and khakis. Business casual became standard by the 2000s. Remote teaching arrived next, showing only shoulders on camera. Students now attend lab in shorts and sandals, and nobody reacts.

Michael Girdley made a similar case on LinkedIn recently. He described clothing as a psychological lever and recommended sitting one notch above whatever baseline a room sets. A classroom's baseline shifted from suits to slacks to jeans across four decades. A professor one notch above jeans wears a collared shirt. Overdressing past that baseline signals distance from students.

Employers see the same pattern at senior project presentations. A team presenting to an industry panel dresses one notch above the panel's own baseline, and that baseline now runs to polos and jeans. Reading the room correctly frees energy for the work itself.

I still own that tie (and a whole bunch more) from the Microsoft photo. I wore it exactly once there. I never did learn the belt and shoes rule I was supposedly breaking. Thirty years or so in, that problem solved itself: nobody expects the tie anymore and I like my tee shirts, shorts and flip-flops.

Why Quantum Computing Matters, Even If You Never Touch One

I logged into my bank app last week to move some money to pay some bills. A padlock icon appeared next to the address bar. Secure connection. I didn't think about it again. Somewhere, someone already has a copy of that session, sitting on a hard drive, waiting.

That padlock runs on a math problem: factor a huge number into its two prime components. A classical computer needs longer than the age of the universe to do that. A large enough quantum computer, running an algorithm built for exactly this job, does it in hours. That computer does not exist yet. It is being built right now, funded by billions of dollars, and no government or company controlling that funding will announce the day it works.

You do not get to wait for that announcement. Security researchers call the attack already underway harvest now, decrypt later. Think about what you have touched on the web this year: your bank, your medical portal, your voter registration, your immigration or legal filings, your tax return, your cloud photo backup. State-backed intelligence services and criminal groups are copying that same traffic from millions of other people today, and storing it cheaply. Almost none of it needs to be cracked now. It needs to still exist, on somebody's drive, the day a quantum computer catches up. Then all of it becomes readable at once.

NIST finalized three new encryption standards built to resist this and set 2035 as the deadline to retire the vulnerable ones. Cloudflare and Google are not waiting until then; they have committed to migrating by 2029. Banks, hospitals, and government agencies holding decades-long records are moving faster still, because for them the clock already ran out on some of what they are protecting.

The fix is not a patch on the old method. It replaces the math underneath it. Government researchers finalized a new set of encryption methods built to survive a quantum computer, the same way today's methods survive an ordinary one. Companies are not waiting to switch over all at once, either. Cloudflare, Google, and Apple already run two locks on the same connection at the same time: the old one and the new one. If someone finds a weakness in the new lock, the old one still holds. If a quantum computer breaks the old lock, the new one still holds. Traffic protected this way stops being worth harvesting, because breaking one lock alone gets an attacker nothing.

Getting every bank, hospital, and government agency onto the new locks is the slow part. It means finding every place the old encryption sits inside a system and replacing it without breaking what depends on it. The federal government has published a shared plan for doing exactly that, across every agency and industry. None of it requires anything from you. It happens inside the apps and websites you already use.

The fix only protects what gets encrypted after it is installed, though. Anything copied under the old lock before that happens is already sitting on somebody's drive, and nothing reverses that. New encryption prevents future harvesting. It does not undo what has already been taken.

Diagram claude.ai generated

Nothing solves the data that is already sitting on somebody's drive. There is no way to reach into another party's storage, revoke a copy, or make it unreadable again. The new locks protect what gets encrypted after they are installed. They have no effect on a copy that left your device years earlier.

How much that matters depends on what kind of data it is. A password can be changed after the fact, so a stolen password loses most of its value once you reset it. A medical record, a Social Security number, or a biometric scan cannot be changed. Whatever gets exposed on that front stays exposed for good. It also depends on how long the data needed to stay private in the first place. Something that only mattered for a few years is probably already safe by the time a quantum computer shows up. Something that needed to stay private for decades, a government file, a company's trade secrets, a hospital's records, is running on a clock that started the day it was copied, not the day the quantum computer arrives.

Most people are not personally worth the effort. A patient attacker spends storage on high-value targets: government communications, corporate research, hospital systems, banks. If you end up exposed, it is more likely through one of those institutions holding your data than through anyone singling you out.

There is not much a consumer can do about data that is already gone. A few things still help:

•   Rotate what can be rotated. Change passwords regularly. A password manager makes it cheap enough to do more than once.

•   Turn on two-factor authentication. It stops a harvested password from being enough to log in on its own, since an attacker also needs your phone, an authenticator app, or a security key. It does not protect the data itself. A hospital record, a legal filing, or an old email that already left your device is untouched by it, because it only fires at login, not on the traffic that carries the data.

•   Not every two-factor method is equal. A text message code can be intercepted or rerouted if someone tricks your phone carrier into moving your number to a new SIM. An authenticator app is safer, since the code generates on your phone instead of traveling over the phone network. A hardware security key is safer still, a small physical device that checks a site's real address before it responds, so it does not work on a convincing fake login page. Use one for your email, your bank, and your password manager itself.

•   Assume what cannot be rotated is already out. Watch for misuse instead of trying to prevent something that may have already happened. A credit freeze and fraud alerts catch someone using a stolen Social Security number long before you would otherwise notice.

•   Keep your software current. The new locks only work if your browser, phone, and apps are recent enough to use them. An old browser is still running the old lock alone.

•   Think twice about what you put online today. Anything sent through email, cloud storage, or a messaging app now can still be harvested under the old lock until your provider finishes switching over.

•   Ask the institutions holding your data. A bank or hospital's own migration timeline affects your exposure more than anything you do personally, and it is a fair question to put to them directly.

None of it undoes what has already been copied. It limits what is still worth harvesting and reduces the damage from what is not.

The padlock on my banking app will look exactly the same through all of this. Behind it, the locks are being swapped one at a time: bank by bank, cloud provider by cloud provider. The question left is not whether the new lock exists. It is how much of your data got copied before it arrived.

Where the Jobs Actually Are

Some graduates spent the 2026 commencement season blaming AI for a job market that shut them out, loud enough that tech executives got booed at graduation ceremonies over it. Recruiters tell a different story.

Matt Walsh, CEO of the Phoenix search firm Blue Signal, works semiconductor hiring daily and says the problem isn't automation. "There aren't enough people," he says. The United States is heading toward what labor economists call the largest workforce shortage in its history, and it shows up hardest in the fields that build things.

The semiconductor industry expects to add close to 115,000 jobs by 2030. The Semiconductor Industry Association projects a shortfall of 67,000 technicians and engineers to fill them. That gap sits squarely in associate degree and bachelor's degree technical programs, not in the AI research labs getting most of the headlines.

Construction and the skilled trades show the same pattern. Branka Minic, CEO of the Building Talent Foundation, says fewer than half the workers needed in construction are entering the field, even with starting wages hitting $50 an hour in some markets. College graduates aren't matching that pay in comparable years of training.

Cybersecurity tells a similar story. CyberSeek, the workforce tracker built by CompTIA and NIST, counts hundreds of thousands of open cybersecurity positions in the U.S. against a supply of qualified workers that consistently falls short. The roles span network defense, security operations, and incident response, and they don't require a computer science PhD. A two-year degree with the right certifications gets a candidate into the field.

This is good news if you're building a technical career instead of chasing a headline. Employers in semiconductors, cybersecurity, advanced manufacturing, robotics, and skilled trades are competing for candidates, not filtering through thousands of applicants for one opening. Two-year technical programs, apprenticeships, internships and engineering degrees put graduates directly into that competition.

States have noticed too: several are merging workforce and higher education agencies or offering loan payoff incentives to pull people into these pipelines.

The AI panic makes for a cleaner headline than a demographic and skills pipeline problem. But the demand for people who can build, install, test, and maintain physical systems is not shrinking. It's the part of the labor market with the fewest applicants and the most openings.

The Free Path Into Quantum Work

The entire IBM Quantum Learning catalog is now open to the public, no partner account or IBM Quantum Network membership required. The catalog runs past ten courses, from the basics of qubits and circuits through algorithms for factoring and search, up to a course built around running experiments on processors with 100 or more qubits, all hosted free on the IBM Quantum Platform.

Path from free IBM Quantum Learning courses to certification to a quantum workforce role

The IBM Certified Quantum Computation using Qiskit v2.X Developer - Associate is a single exam built on the same Qiskit SDK and Qiskit Runtime the free courses teach. Since the program launched in 2021, more than 1,300 people across 71 countries have passed it. The exam runs up to $200 depending on location, with an optional $30 practice test through Pearson VUE. The preparation runs $0.

This matters most for programs without a quantum lab or a research grant behind them, which describes most community colleges and a fair number of universities. Qiskit needs Python and a working knowledge of linear algebra. Nothing else.

Every one of my students has a laptop and an internet connection. That's all you need to get started today.

This post rolls into the next edition of Quantum from the Ground Up, due September 1.

The Quantum Chandelier

Photo MIT Technology Review

You've probably seen a photo like this - a tower of gold-plated discs wrapped in loops of wire, narrowing toward a point at the bottom. Google, IBM, and most quantum computing press releases use some version of this image. People call the assembly the chandelier, and it looks like the computer. It isn't. It's the life support system for a chip you can't even see in the photo.

Start with what's actually inside a quantum computer. A qubit is a physical device, usually a tiny loop of superconducting metal, that can hold a mix of two states at once instead of a single 0 or 1. That mixed state is fragile. A stray photon, a vibration, or a few millikelvin of extra heat can collapse it before you get a useful calculation out of it. Physicists call that collapse decoherence, and it's the central engineering problem in the entire field. Keeping decoherence at bay is the whole reason the chandelier exists.

Why It Has to Be That Cold

At room temperature, everything around a chip is radiating heat as stray photons, trillions of them, bouncing around and hitting anything nearby. For a normal computer chip that's irrelevant. For a superconducting qubit it's fatal, since a single one of those stray photons carries enough energy to flip the qubit's state. Cooling the chip down to the mixing chamber stage, near 10 millikelvin starves the environment of those stray photons and also lets the qubit's own wiring become superconducting, meaning it carries current with zero electrical resistance. Both effects are required. Without one or the other, the qubit decoheres in nanoseconds instead of the few hundred microseconds researchers need.

The Five Stages

The gold structure is a dilution refrigerator, built from stacked stages that grow colder toward the bottom. A pulse tube cryocooler, essentially a specialized mechanical compressor, does the first heavy lifting, dropping the system from room temperature to about 40 Kelvin and then 4 Kelvin using compressed helium gas. Below that, the fridge switches to a different method. A chamber called the still boils off helium-3 to reach roughly 0.7 Kelvin, and a series of heat exchangers pushes the temperature down further, through about 0.1 Kelvin, to the mixing chamber at the very bottom. A mixture of helium-3 and helium-4 drives that last stage. The full cooldown from room temperature to base takes 24 to 48 hours, and it has to happen every time the system needs to be opened for maintenance.

Each gold disc in the photo is one of these stages, plated in gold because gold conducts heat well, resists corrosion, and doesn't interfere magnetically with the qubits. Each stage nests inside the next, shielding the colder one below it from the warmer one above.

Figure: the chandelier's cooling stages, colored from gold at the warmest to blue at the coldest, with the qubit chip mounted at the base.

The Wiring Problem

Every wire running through those stages carries control signals down to the qubits and readout signals back up to room-temperature electronics. A signal heading down gets attenuated at each stage on the way, stripping out electrical noise picked up from the warmer stages above. A signal heading up gets amplified, since the qubit's own readout signal is too faint to detect at room temperature. Every one of those wires is also a heat leak. Heat travels down a wire just as easily as a signal does, and the cooling power at the base stage is measured in microwatts, barely enough to warm a fraction of a grain of rice. A single wire that's improperly thermalized can add more heat than the entire fridge can remove.

That tradeoff is why wiring, not the qubits themselves, has become one of the field's biggest scaling obstacles. Each qubit needs its own set of control and readout lines, so the number of cables required grows directly with qubit count, eventually exceeding what a single cryostat can physically hold. Labs are now racing to move some control electronics inside the fridge itself, onto chips that can survive the cold, so fewer wires have to cross from room temperature down to the millikelvin stage. A single dilution refrigerator system runs one to five million dollars, and most of that cost is solving this wiring problem.

What's Actually Quantum

The qubit chip itself sits at the very bottom, bolted to the mixing chamber, a few millimeters across. It's small enough to miss in most photos of the chandelier, which is the point. Everything above it, every gold disc, every coil of coax, every attenuator and amplifier, exists for one reason: to keep that small chip cold and quiet enough to hold a quantum state long enough to be useful. Chapter 3 of Quantum from the Ground Up makes this argument directly: physics works, engineering makes it work. The chandelier is that argument built out of gold-plated copper and helium-3.

Next time a chandelier photo turns up in your feed, look past the wiring for the small chip at the bottom. That chip is the entire quantum computer. Everything else in the picture is plumbing.

This post will fold into the next edition of Quantum from the Ground Up, out September 1.

The Second June 22, 2026 Quantum Executive Order

In a post Tuesday, I covered the executive order setting hard federal deadlines for post-quantum cryptography migration: key establishment by 2030, digital signatures by 2031. That order dealt with defense. On the same day, June 22, the president signed a second order that deals with offense.

Executive Order 14413, "Ushering in the Next Frontier of Quantum Innovation," directs the federal government to build a large-scale quantum computer for scientific use. The centerpiece is the Quantum Computer for Application Development and Discovery Science effort, called QC-ADDS. The order directs the Department of Energy to deliver at least one QC-ADDS system to a DOE facility and make it available to the scientific community.

Here's some details - within 90 days, DOE must publish the technical specifications required for QC-ADDS to perform transformative scientific applications beyond current classical computer capabilities. Within 180 days, DOE must explore private-sector partnership models and report on cost, scope, and delivery timeframe. DOE has already responded: its Quantum Genesis initiative targets a fault-tolerant, scientifically relevant quantum computing capability by 2028, with a National Quantum Supercomputing User Facility to give U.S. researchers access to systems across multiple qubit modalities.

The Commerce Department must develop a plan for advance market commitments to pull in commercial quantum vendors. The Defense Department gets its own track, establishing programs for national security applications of quantum computing, potentially including a dedicated center. The order also establishes a national center for quantum performance assessment and directs a government-wide quantum workforce recruitment strategy, including special pay rates and retention incentives.

The workforce section carries the most direct relevance for technical education programs. The order tasks NSF to stand up a network of National QIST Workforce Development Institutes within 180 days. Federal money for hands-on QIST training will flow somewhere; the question is where.

There is a thread connecting both orders. The PQC migration order sets a deadline for protecting existing systems. EO 14413 sets a timeline for building the systems that will eventually make those protections necessary. Both orders treat 2030 as the planning horizon. Harvard's Mikhail Lukin put fault-tolerant, large-scale quantum computers at end-of-decade in a recent assessment, five to ten years ahead of earlier estimates.

Government Sets New Deadline for Quantum-Safe Encryption

A student in one of my summer courses asked the question I get every time encryption comes up in discussion: why does this matter now? RSA (Rivest-Shamir-Adleman) and ECC (elliptic curve cryptography) have protected data for decades. The quantum computer that breaks them does not exist yet. 

My usual answer leans on Q-Day estimates: Google's Gidney put the threshold at roughly one million physical qubits to break RSA-2048, and an IonQ fidelity result last October pushed the realistic window to somewhere between 2029 and 2033. Most expert estimates before that sat closer to 2035. On June 22, the federal government answered the student's question for me. President Trump signed 

an executive order setting hard deadlines for federal post-quantum cryptography migration (PQC): agencies must move high value assets to post-quantum key establishment by December 31, 2030, and post-quantum digital signatures by December 31, 2031. Federal contractors get the same 2030 deadline for FIPS (Federal Information Processing Standards) compliance.

That replaces the prior government baseline. Under the Biden administration's National Security Memorandum 10, agencies were planning around 2035. The new order compresses that by four to five years and adds teeth: agencies must name a PQC migration lead within 30 days, the Commerce Department must run a migration pilot by the end of 2027, and contractors face FIPS enforcement through procurement rules. 

Coverage from Cybersecurity Dive notes the order also pushes CISA (the Cybersecurity and Infrastructure Security Agency) to publish guidance on cryptographic bills of materials, the inventory work agencies need before they can migrate anything.

How the Industry Responded

Two days after the signing, STMicroelectronics introduced the ST54M, the first mobile chip with a dedicated hardware accelerator for post-quantum algorithms. It runs ML-KEM (Module-Lattice-Based Key-Encapsulation Mechanism) and ML-DSA (Module-Lattice-Based Digital Signature Algorithm), the NIST (National Institute of Standards and Technology) standards finalized in 2024, on a single die alongside NFC (near-field communication), secure element, and eSIM (embedded SIM) functions. Commercial sampling is available now, with certification targeted for July 2026. That is the hardware path the federal order is pushing the rest of industry toward on the same compressed timeline.

I tell students today: nobody knows the exact day a cryptographically relevant quantum computer arrives, but the government just stopped waiting to find out. And.... I would not be surprised at all to see the deadline moved forward again.... soon.

STEM at Two Years: Community College Degrees That Pay

Most of my career has been at the community college. I directed an NSF Center of Excellence at Springfield Technical Community College and taught electronics, computer systems, and photonics there. At Holyoke Community College I still teach engineering transfer courses part time for students heading to four-year universities. Over forty years I have watched students come through two-year STEM programs and go directly into careers that surprised people who assumed a bachelor's degree was required. This post is the third in a series on degree choice and outcomes. The first two covered bachelor's programs and two-year degrees broadly. This one focuses specifically on STEM at the associate degree level: what the programs are, what they pay, and how the job outlook looks in 2026.

The macro case for STEM at any credential level is straightforward. The BLS projects STEM occupations will grow 8.1 percent between 2024 and 2034, nearly triple the 2.7 percent rate for all other occupations. The median salary across STEM occupations sits at $101,600, well above the all-occupation median. The two-year credential does not open every STEM door, but it opens more of them than most people expect, and it does so at a fraction of the cost and time of a four-year path.

The highest-paying two-year STEM programs in 2026, per BLS occupational data: information security analysts (cybersecurity) median at $119,860 with 32 percent projected job growth through 2032; radiation therapy at a median above $100,000; dental hygiene at $94,260; and registered nursing at $93,600. Below those, nuclear technicians median around $84,000, electronics engineering technicians around $67,550, and laser electro-optics technicians in the $55,000 to $65,000 range depending on industry and region. HVAC technology and computer network support round out the middle of the table at $58,000 to $62,000.

A point worth making clearly: the two-year STEM credential typically leads to technician and support roles, not engineering or research positions. That distinction matters for career planning, but it does not diminish the outcomes. An electronics engineering technician working in manufacturing or test and measurement earns $67,550 median with stable demand. A cybersecurity analyst with an associate degree and relevant certifications, CompTIA Security+ in particular, enters a field with 32 percent projected growth and a six-figure median salary. The ceiling in those careers depends more on certification, experience, and specialization than on whether the entry credential was a two-year or four-year degree.

The cost side of this decision matters as much as the salary side. Average annual tuition at a public two-year college runs about $3,990, versus over $11,500 at a public four-year institution. A student completing a two-year cybersecurity or nursing program graduates with little or no debt and enters a field paying $90,000 to $120,000. A student completing a four-year program in the same field earns more in some cases, but starts with average student loan debt above $29,000 and two additional years of foregone income. For STEM technician roles specifically, that math favors the two-year path more consistently than in most other fields.

Before committing to a two-year STEM program, check three things. First, verify that the program carries the right accreditation for your field. Nursing programs must be accredited by ACEN or CCNE for graduates to sit for the NCLEX. Engineering technology programs are credentialed by ABET. Second, check whether the career path requires licensure or certification beyond the degree itself, and build the cost and timeline for those credentials into your plan. Third, look at your specific college's job placement data for that program. National medians are a baseline; local labor market conditions move those numbers significantly in both directions.

One pathway that gets less attention than it deserves: the two-year degree as the first half of a four-year degree, paid for by an employer. Many community college STEM graduates enter the workforce directly, then pursue a bachelor's degree part time while their employer covers tuition. This is not rare. A significant share of working adults completing bachelor's degrees are doing exactly this, particularly in nursing, engineering technology, and information technology. The RN-to-BSN pathway is the most established example: a graduate earns an associate degree, passes the certification, enters the workforce as a registered nurse, and completes a BSN online or part time over two to three years, often with hospital tuition reimbursement covering most of the cost. The same model applies in engineering technology and cybersecurity, where employers in manufacturing, defense, and infrastructure actively fund continuing education. The credential upgrade from technician to technologist, meaning from associate to bachelor's degree, also typically comes with a pay bump and expanded career options. For students weighing cost, this route splits the financial risk: two years of low-cost community college tuition, then employer-subsidized completion of the bachelor's, with income throughout. The total credential is the same four-year degree. The debt load and the timeline are very different.

The community college students I’ve watched who did best in two-year STEM programs were not picking a fallback. They were picking a specific job in a specific field and treating the degree as the direct path to it. That approach still works in 2026. For some, the two-year degree is also the starting point for a four-year degree the employer ends up paying for. The programs are there. The jobs are there. Check the current numbers before you decide. Know the program, know the credential requirements, know the market.

Both Tracks Moving

In 1994 I started writing a textbook called Windows 95 Essentials for an Engineer’s Toolkit. There was one problem: Windows 95 did not exist yet. Microsoft was still building it, and they pushed updates almost weekly. Each one arrived on a new set of over a dozen floppy disks. Every update meant loading those disks, reinstalling from scratch, retesting every procedure, and rewriting any section that no longer matched the software. My second daughter, Gabby was turning four years old. Eva was born in June 1995, right as the book was finishing. I kept writing.

Spring 1995 something else happened that had nothing to do with floppy disks. The internet was being privatized in real time. Through most of the early 1990s, the internet was a government and academic network. The NSFNET backbone carried U.S. research and education traffic at no cost to institutions. Commercial access was limited, and online services like CompuServe and AOL operated as walled gardens: you paid a subscription, you got their content, and the wider internet was largely off the table. Microsoft had built The Microsoft Network on exactly that model, a paid subscription service meant to compete with AOL. Then the walls started coming down. The NSFNET was decommissioned in April 1995, handing the backbone to commercial providers. Commercial ISPs multiplied. The web browser arrived. And Microsoft, watching the same thing everyone else was watching, pivoted almost overnight. Bill Gates’ “Internet Tidal Wave” memo from May 1995 called the internet “the most important single development to come along since the IBM PC.” MSN shifted toward the open web. 

Technically, Internet Explorer (IE) did not ship with the OS on August 24,1995 - IE 1.0 was released a week earlier on August 16,1995 as part of what Microsoft called the Plus! pack.

The book I was writing had to reflect a platform that was no longer just a desktop operating system; it was suddenly a node on a network becoming public infrastructure. That meant more rewrites. It also meant the book was documenting something larger than a software release.

The book took nearly a year to complete. The only way through it was parallel progress. I could not wait for the software to stabilize before writing, and I could not wait for the writing to be done before testing. Both tracks ran at the same time, and I updated whichever one had fallen behind. That is not a comfortable way to work. It is, however, the only way to finish something when the target keeps moving.

Working under moving targets is a skill. Most engineering projects involve some version of it: a component spec changes mid-build, a client requirement shifts after the design review, a test result forces a redesign two weeks before the deadline. The teams that handle this well are not the ones with the most complete plans. They are the ones who keep both tracks moving and update each one as new information arrives.

Three practical habits help. First, document as you go rather than saving it for the end; late-stage documentation of early decisions is mostly reconstruction from memory. Second, treat a changing spec as new information, not a setback; the project is not broken, it has just been updated. Third, keep the physical work and the written work in sync; a prototype that is ahead of its documentation, or documentation that describes a prototype that does not exist yet, creates debt that compounds.

The floppy disks eventually stopped coming. The book shipped. Eva was born healthy. Windows 95 launched on August 24, 1995, with a Rolling Stones song and more press coverage than any software release before it. The 31st anniversary is two months away. Not a round number, but the lessons from that year still hold: keep both tracks moving, treat every spec change as information, and do not wait for conditions to settle before making progress. The target never stops moving.