AI History
In 1950, a British mathematician named Alan Turing asked a question that changed everything: Can machines think?
It wasn't a rhetorical question. Turing proposed an actual test - could a machine hold a conversation indistinguishable from a human? At the time, there were no computers powerful enough to attempt it. But the question itself launched a field.
Six years later, at a summer conference at Dartmouth College in 1956, John McCarthy and a group of researchers officially coined the term "artificial intelligence" and established it as a formal discipline. The goal was ambitious: describe human intelligence precisely enough that a machine could replicate it. That goal proved far harder than anyone expected. What followed was seventy five years of breakthroughs, dead ends, and resets.
Rule-Based Logic (1950s–1970s)
Early AI ran on if-then rules. Programs could play chess, solve equations, and simulate basic conversation, but every possible scenario had to be manually coded in advance. It couldn't adapt. It couldn't learn. And when researchers promised more than the technology could deliver, funding dried up in the 1970s during the first of two periods known as AI winters - stretches of stagnation caused by unmet expectations and major cuts in research investment.
In buildings: The same rule-based logic powered early HVAC controllers. Set point reached, system responds. Simple, effective, and completely rigid.
Machine Learning (1980s–2000s)
Researchers stopped coding rules and started teaching machines to find patterns in data. Feed the system examples. Let it figure out the rest.
By the 1990s, real results emerged. In 1997, IBM's Deep Blue defeated world chess champion Garry Kasparov - a landmark moment that demonstrated AI's growing capability to outperform humans in specific intellectual tasks. The critical ingredient was data. More of it meant better decisions.
In buildings: BMS platforms matured. Trends and remote monitoring entered the industry. Buildings started generating operational data. Most of it went unused, but the foundation was being built.
Deep Learning (2010s)
In 2012, a neural network called AlexNet shattered image recognition benchmarks at the ImageNet competition. Developed by Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton, it demonstrated that large scale neural networks trained on GPUs could outperform every previous model by a wide margin.
The techniques and momentum generated by AlexNet helped shape the development of modern natural language processing models that power tools like ChatGPT. Google's 2017 Transformer paper became the technical backbone of every large language model that exists today.
In buildings: AI powered BMS began learning from operational data - predicting equipment failures, optimizing energy without manual reprogramming, adapting to occupancy in real time. Predictive maintenance moved from theory to field deployment. The tools existed. Adoption was still uneven.
Generative AI (2020s–Present)
On November 30, 2022, ChatGPT launched. Within five days it had one million users. By January 2023, it had surpassed 100 million - making it the fastest-growing consumer application in history. TikTok took nine months to reach the same milestone. Instagram took two and a half years.
For the first time, AI could hold a real conversation, write a coherent document, explain a complex problem, and adapt to follow-up questions - in plain language, with no technical knowledge required.
In buildings: The gap between technology and field application has never been smaller. Property managers and chief engineers now have direct access to tools that can draft tenant communications, generate SOPs, build inspection checklists, and interpret building data - in minutes, from a phone.
What Seventy Five Years Actually Produced
Most of the researchers who launched this field in 1956 were trying to replicate human intelligence. They didn't get there. What they built instead, after decades of false starts, funding collapses, and unexpected breakthroughs was something more useful: tools that can sit beside an experienced leaders and make their documentation tighter, their decisions faster, and their team more effective.
The technology didn't replace judgment. It got good enough to amplify it. That's the opening CRE has been waiting for.
CRE Pro: What This History Made Possible
This industry has a knowledge transfer problem that shows up every single day. People walk into property management roles with real responsibility - buildings, tenants, vendors, emergencies and the operational knowledge they need is almost never written down anywhere. It lives in the heads of experienced leaders, passed on informally, or learned the hard way through incidents that shouldn't have happened.
CRE Pro is an AI tool I built specifically for CRE operations. A tool calibrated for the work property managers actually do - incident response, SOP development, vendor oversight, tenant communication, budget preparation - designed to meet you in the moment you need it.
Ask it:
Help me learn about XYZ building systems and equipment.
Help me build a vendor performance checklist for XYZ service.
What should I be thinking about during emergency response that most PMs miss?
I'm preparing a calendar for tenant events, what are some ideas for the month of XYZ?
What should I be doing in the first 90 days at a new building?
Draft a tenant communication for planned elevator work next Friday.
Draft a walkthrough checklist for a industrial property.
Pin it to your ChatGPT sidebar. Use it before your next tough conversation, your next incident, your next SOP that's been sitting on the to do list for six months. Treat it like a mentor who has managed buildings, knows the vocabulary, and is available at 6am when the panel activates.
Let me know in the comments what you asked it!
Remember, always follow your company and owner AI policies.
