Every technology revolution feels unprecedented. I’ve lived through four of them. Here’s what actually works.
The Panic is Understandable But Misplaced
I know that I have felt it. When a particularly aggressive shift happens. If you’re anxious about AI transforming your career, that’s rational. You’re watching:
- ChatGPT write code in seconds that took you hours
- Salesforce/AWS deploy AI agents to 12,000 companies overnight. Adding Snowflake/Anthropic? That’s 24,600.
- Job postings requiring “AI proficiency” for roles that didn’t mention it six months ago
- Your company announcing “AI transformation” with no clarity about what happens to your role
You’re thinking: “This time is different. AI is changing EVERYTHING. I can’t keep up!”
I’m here to tell you: Every technology revolution felt exactly like this.
Historical reality. Every generation said the same thing:
- 1980s: “Computers will eliminate all jobs”
- 1990s: “The internet changes everything”
- 2000s: “Outsourcing will destroy tech careers”
- 2010s: “Cloud computing eliminates IT jobs”
- 2020s: “AI will replace knowledge workers”
What actually happened: Technology eliminated specific tasks. It created demand for people who could deploy it effectively in real-world contexts.
The workers who thrived across all those transformations:
- Built foundation skills that translated across shifts
- Expected continuous learning (didn’t panic at refresh cycles)
- Combined technical capability with domain expertise
- Positioned as bridges (technical + business understanding)
The workers who struggled:
- Chased specific job titles that became obsolete
- Relied on static credentials (degrees that didn’t update)
- Panicked at each technology shift
- Refused to learn new tools
The Path Forward Isn’t New
When I first started out in the early 1980s, there were no computers in our high school (for context, I graduated in 1982). No cellphones, no pagers. Nothing like that. No one even talked about altnet back then. We had one guy that played around with building robots in his garage. The pinnacle of “modern technology” for us. People – girls primarily – that wanted to go into business were taking typing classes on IBM Selectric typewriters. MACs and PCs weren’t even on our radar let alone Apple brand. I vaguely knew there were “computers” (see EDVAC and ENIAC). Those were outside of my understanding or exposure so shoved aside in my mind.
However. Like most when I graduated, I needed a job.
After a brief stint of restaurant shift work, I signed up with Manpower. I had a lot of different positions, and was successful within their structure. But more importantly, that experience gave me exposure to ideas and new career paths I had not considered.
What worked for me (1984-2025):
- Short-term training (9 month course – Data Processing and Business Systems) → entry role
- Continuous certifications (Netware, Novell, Microsoft, Cisco, AS/400 etc)
- Domain expertise (manufacturing, finance, operations) + technical skills
- Bridge role (translate technical to business needs)
- Adapt every 2-3 years (expect it, don’t fear it)
Result: 40-year career through mainframes, PCs, client-server, internet, cloud, and now AI. Never finished that bachelor’s degree but I have hundreds of credit hours. Completed general requirements then just kept adapting. Foundation skills + continuous learning outlasted formal credentials. Programming. Project management. Adding and changing with the tides.
What works now (2025+):
- Bootcamp/online learning (3-6 months) → entry technical role
- Continuous certifications (Python → AWS → AI platforms)
- Domain expertise (healthcare, finance, manufacturing) + technical literacy
- Bridge role (Forward Deployed Engineer, AI implementation, AI SEC Ops, compliance)
- Adapt every 12-18 months (expect it, don’t fear it)
Same pattern. Faster cycles. Same fundamentals.
How I Know This: The Big Box With Blinking Lights
My second job was at a company that had spent serious money on what they proudly called “their computer system.” It was a big box mainframe with spinning disks and blinking lights with lots of buttons to push that didn’t power much of anything. Not inventory management, not accounting, nothing that would justify the investment. Basically it stored information. No calculations or advanced data management. Just storing information.
It sat there. Expensive. Impressive-looking. Essentially useless by today’s standards.
My first job was at a school and wasn’t much better. They’d bought a computer system and used it as a fancy word processor for mail merge letters. Thousands of dollars in hardware to do what a typewriter and a mailing list could handle. It was even placed in the mailroom. Names, addresses, and a monthly newsletter was it’s rai·son d’ê·tre. My job was to make sure that any address changes were entered into the system and if alumni had died, update their status. Then, once a month, print out the address labels for the newsletter. Yes, seriously.
I was their first computer operator. I’d taken that nine-month course by this time. More training than anyone else at the school had with computers that weren’t mainframes with punch cards. That made me the expert by default.
These companies, and others I worked for later, had no idea what to do with the technology. Neither did I, really. We were all figuring it out together.
Sound familiar?
Forty years later, I’m watching companies spend billions on AI infrastructure they don’t know how to use. Salesforce Agentforce deployed to 12,000+ enterprise customers who’ll implement it because it’s available, not because they’ve figured out how it fits their operations. Data centers getting built – literally more big boxes with blinking lights – processing AI workloads that haven’t translated to clear business value.
I’ve seen this before.
The Pattern That Keeps Repeating
Phase 1: Infrastructure Before Application
1980s: Companies bought computers before knowing what to do with them
- Expensive equipment sitting underutilized
- “We need this technology” without clear use cases
- Early adopters figuring it out by trial and error
2020s: Companies buying AI capability before knowing how to apply it
- Salesforce Agentforce deployed to thousands who’ll underutilize it
- “We need AI strategy” without clear implementation plans
- Forward Deployed Engineers figuring it out in real customer environments
Phase 2: Formal Education Can’t Keep Pace
1980s-90s: Universities teaching mainframe programming while PCs took over
- Curriculum committees moving too slowly
- By the time requirements updated, technology had moved again
- Computer Science degrees teaching theory disconnected from practice
2020s: Universities teaching traditional software development while AI automates coding
- CS programs still structured around pre-LLM assumptions
- “Learn to code” advice outdated by the time students graduate
- Coding bootcamps struggling to stay relevant
Phase 3: Certifications Replace Degrees
1980s-90s: Vendor certifications became currency
- Netware certification > Computer Science degree for network jobs
- Microsoft certification > formal education for systems administration
- Cisco certification > university networking courses
2020s: AI platform certifications emerging
- AWS AI certifications
- Google Cloud ML Engineer
- Anthropic Claude implementation
- Not about proving you “know” the tool, but that you can learn it
Phase 4: Refresh Cycles Accelerate
1980s-90s: Technology refresh every 2-3 years
- Netware 3.x → 4.x → 5.x (major architectural changes)
- Windows NT → 2000 → XP → Vista (constant reinvention)
- Cisco IOS updates requiring recertification
2020s: AI refresh every 12-18 months
- GPT-3 → GPT-4 → o1 → GPT-5 (rapid capability jumps)
- Claude 1 → 2 → 3 → 3.5 → Opus 4 (continuous iteration)
- Agentforce 1.0 → 2.0 in months, not years
Phase 5: The Bridge Role Emerges
1980s-90s: Someone had to translate technology to business needs
- I was the “Forward Deployed Engineer” of 1984
- We were making computers work in specific business contexts
- Every implementation was custom (no templates, no playbooks)
- Roles evolved: computer operator → systems administrator → network engineer → architect
2020s: Forward Deployed Engineers doing the same job
- Making AI agents work in specific customer environments
- Every deployment custom (platforms handle 70%, humans handle complex 30%)
- Role will evolve: FDE → whatever comes next
- But the function (bridge technical + business) remains
What Actually Worked (Then and Now)
1. Foundation Skills > Formal Credentials
What I learned:
- Nine-month training course > incomplete bachelor’s degree
- Understanding systems thinking > memorizing syntax
- Troubleshooting methodology > specific technical knowledge
- Continuous learning capability > static expertise
What this means now:
- Python bootcamp + domain expertise > CS degree without specialization
- Understanding how AI systems fail > knowing how to prompt ChatGPT
- Problem diagnosis skills > tool-specific knowledge
- Learning agility > current skill stack
2. Domain Expertise Is Portable
What I learned:
- My understanding of business operations grew with each changed position and translated across every technology shift
- Manufacturing domain knowledge worked whether the system was mainframe or client-server
- Financial operations expertise applied to DOS, Windows, cloud
- The domain stayed constant; the tools changed
What this means now:
- Healthcare coordinator’s hospital operations knowledge translates across AI tools
- Manufacturing expertise applies whether you’re working with legacy systems or AI agents
- Financial regulations knowledge works with spreadsheets or synthetic data generation
- Domain expertise outlasts technology platforms
3. Certifications Prove Learning Ability
What I learned:
- Netware certification didn’t mean “I’ll know Netware forever”
- It meant “I can learn complex systems and apply them”
- Every recertification proved I could adapt
- The certifications weren’t the endpoint—learning capability was
What this means now:
- AWS AI certification proves you can learn new platforms
- Python proficiency proves you can acquire technical skills
- Each new tool mastered demonstrates adaptation capability
- Certifications are proof of learning ability, not static knowledge
4. Expect Constant Reinvention
What I learned:
- 2-3 year refresh cycles weren’t a bug, they were the feature
- Technology evolution was the constant
- Workers who panicked at each shift burned out
- Workers who expected change thrived
What this means now:
- 12-18 month AI refresh cycles are the new normal
- Models will improve, tools will change, platforms will consolidate
- Workers panicking about “AI moving too fast” need perspective
- This IS how technology transformation works. Always has been
5. Be the Bridge
What I learned:
- Technical capability is commodity (everyone eventually learns the tools)
- Translating technical to business is scarce (requires both domains)
- I succeeded because I could explain system capabilities to manufacturing managers
- That translation skill – human context + technical understanding – never automated
What this means now:
- Coding is increasingly automated (GitHub Copilot, Claude Code, vibe coding)
- Explaining technical trade-offs to executives or the workforce isn’t (requires human judgment)
- Forward Deployed Engineers succeed by bridging AI capability and business reality
- That translation skill remains human territory
The Foundation Skills That Worked Then, Work Now
My 1984 Stack → 2025 Equivalent
| Then (1984-2005) | Now (2025+) | Why It Translates |
| Netware/Novell networking | Cloud infrastructure (AWS/Azure/GCP) | Systems integration thinking |
| Microsoft enterprise systems | Enterprise platforms (Salesforce/Snowflake) | Business application logic |
| Cisco routing/switching | API integration, systems architecture | How components communicate |
| Troubleshooting methodology | Problem diagnosis (technical + business) | Human judgment, pattern recognition |
| Documentation/audit trails | Governance, compliance, oversight | Proving decisions were made correctly |
| Stakeholder communication | Translation (technical ↔ business) | Bridge building (can’t be automated) |
The constant across 40 years:
- Systems thinking (how components interact)
- Troubleshooting (what went wrong, why, how to fix)
- Business context (what actually matters to operations)
- Continuous learning (expect change, adapt to it)
- Translation (explaining technical to non-technical)
None of those got automated. They’re MORE valuable now.
What We’re Changing at Under the Radar
I’ve been publishing Under the Radar weekly for a few months now, tracking which career opportunities offer protection against AI displacement. We rank the Top 5 positions each week based on automation risk, market demand, and future viability.
But something’s been missing.
We’ve been telling you WHAT jobs to target. We haven’t been teaching you HOW to build career resilience across technology transformations.
That changes now. Push button agents were the catalyst needed to institute that change.
Starting next week, Under the Radar will include a new section: Foundation Skills.
Instead of just listing job titles (which might not exist in 5 years), based purely on the best available data, we’ll show you:
For University Students:
- Which majors create optionality in the AI era
- Course selection strategy (regardless of major)
- Hybrid degrees that combine domain + technical
- What to avoid (degrees obsolete before graduation)
For Career Changers:
- Foundation skills that translate to multiple protected paths
- Compressed learning timelines (6-12 months to repositioning)
- Leveraging existing domain expertise + adding technical literacy
- Certifications that prove learning capability
For Currently Employed:
- Strategic skill building while working
- Which foundations to prioritize based on your field
- Positioning before you need to (while employed = better negotiating position)
- How to identify when it’s time to pivot
The Skills Matrix:
We’ll show which foundation skills map to which opportunities. For example:
- Python appears in 3 of our Top 4 career paths
- Domain expertise protects across all 4
- Governance/compliance thinking opens multiple paths
- Stakeholder management (human translation) can’t be automated
Learn Python + your industry domain + governance frameworks = qualified for multiple protected roles, not just one job title.
Why this matters:
I never finished that bachelor’s degree. But I built foundation skills. Systems thinking, troubleshooting, domain expertise, continuous learning. That worked for 40 years through mainframes, PCs, client-server, internet, cloud, and now AI.
Those foundations didn’t change. The tools did.
If I’d chased specific job titles (“mainframe operator!”), I’d have been obsolete by 1995. Instead, I built transferable capabilities that worked across every transformation.
That’s what we’re going to help teach you now.
The Bottom Line
I was the first computer operator at multiple companies that didn’t know what computers were for. I took a nine-month course when universities couldn’t teach it fast enough. I never finished that bachelor’s degree because requirements kept changing. I built a 40-year career on certifications that refreshed every 2-3 years for the first couple of decades, then I just added skills as required on my own.
And now I’m watching the exact same pattern with AI.
This has happened before.
Technology can be transformative – for companies and for people.
The workers who panicked about “everything changing”? Some adapted and thrived. Others chased static credentials and got left behind.
The degree programs that couldn’t keep up? Universities eventually figured out that teaching thinking > teaching tools.
The pattern always resolves the same way:
- Infrastructure comes first, applications follow
- Formal education lags, practical learning leads
- Certifications prove capability, not endpoints
- Foundation skills outlast specific technologies
- Bridge builders (technical + business) stay employed
If you’re panicking about AI: I’ve been through this four times. It always feels unprecedented. The fundamentals never change.
If you’re a student choosing a major: Don’t bet everything on “Computer Science = safe.” Bet on hybrid skills (domain + technical), governance capabilities, and learning agility.
If you’re a displaced worker: You’re not starting over. You have domain expertise that’s portable. Add technical literacy (3-6 months), and you’re positioned for multiple paths.
If you’re currently employed but worried: Now is the time to build foundations while you have income stability. Python + governance + your existing domain expertise = options when you need them.
Technology revolutions feel unprecedented while you’re in them.
But the pattern is always the same: foundation skills + continuous learning + domain expertise > formal credentials.
I know because I’ve lived through this before.
Next Week in Under the Radar: We’re launching our new Foundation Skills section. We’ll show you which capabilities create career resilience across AI transformation, with specific learning paths for students, career changers, and currently employed workers.
Subscribe: theopenrecordl3c.substack.com
