The AI Echo of 1985: From “Softer Software” to Today’s Copilots
In 2026, artificial intelligence permeates nearly every aspect of the digital world – writing code, generating images, powering customer service. Yet, the anxieties surrounding AI aren’t new. As far back as 1985, industry leaders were already cautioning against the hype and potential misuse of the technology, a sentiment remarkably similar to today’s discourse.
The “Lemming-Like Rush” to AI
Mitch Kapor, chairman of Lotus Development, warned in 1985 of a “lemming-like rush” to artificial intelligence. He saw an opportunity for those who understood AI’s practical applications, even as others chased unrealistic promises. This echoes current concerns about the rapid deployment of generative AI without a full understanding of its capabilities and limitations.
InfoWorld’s editorial director, James E. Fawcette, questioned whether 1985 would be the year AI moved beyond “ivory towers” to turn into a truly useful tool. Software companies, he observed, were seeking a “savior” to revitalize interest in a market saturated with similar products. The core problem, then as now, was defining what AI could realistically achieve.
Two Faces of AI Abuse: Hype and Overdesign
Fawcette identified two key pitfalls: “AI-hype” – programs promising to run your business with minimal input – and “Rube Goldberg overdesign” – overly complex systems attempting to solve grand, impractical problems. These patterns are readily apparent today, from startups promising fully autonomous business operations to overly engineered AI solutions that fail to address specific needs.
Pro Tip: Before investing in any AI solution, clearly define the problem you’re trying to solve and assess whether AI is the most appropriate tool. Avoid solutions that promise too much without demonstrable results.
Bill Gates’ “Softer Software” Vision
Bill Gates, recognizing the potential for misinterpretation, proposed “softer software” as a more grounded approach. This concept, explored with Charles Simonyi in 1983, focused on systems that learned from user behavior and adapted to individual needs. Simonyi described software that would “modify its behavior over time, based on its experience with the user,” aiming to simplify tasks in the “real world.”
This vision of “softer software” foreshadows today’s personalization engines and adaptive AI copilots. The idea was to create systems that assisted users, rather than attempting to replace them with “thinking machines” and the associated “Large Brother” imagery.
Excel: A Practical Application of “Softer Software”
Microsoft’s early “expert systems” for Multiplan, the precursor to Excel, exemplified this approach. These systems could build formulas, categorize data, and analyze information, making spreadsheets more accessible and powerful. Excel’s “learn-by-example” macro feature, introduced in 1985, allowed users to automate tasks without needing programming knowledge.
Excel’s success wasn’t based on magical promises, but on delivering “consistency, power, lots of features, and macros.” It prioritized practical functionality over grandiose claims, a lesson still relevant today.
From Oracle to Copilot: The Evolving Role of AI
The 1985 InfoWorld editorial envisioned software that could generate reports and charts automatically. This capability is now commonplace, with AI systems routinely drafting documents, summarizing meetings, and creating visualizations from simple prompts. These systems function as “agents,” executing multi-step tasks across various applications.
The key takeaway from this historical perspective is the importance of viewing AI as a copilot, not an oracle. Bill Gates didn’t dismiss intelligence in software, but he rejected the mythology surrounding it. The focus should be on creating systems that learn, adapt, and assist users, rather than attempting to replicate human intelligence.
FAQ: AI Then and Now
Q: What was “softer software”?
A: A concept proposed by Bill Gates and Charles Simonyi as a more realistic goal than “artificial intelligence,” focusing on systems that learn from user behavior and adapt to individual needs.
Q: What were the two main concerns about AI in 1985?
A: AI-hype (programs promising unrealistic automation) and Rube Goldberg overdesign (overly complex systems solving impractical problems).
Q: How does Excel relate to the “softer software” concept?
A: Excel’s features, like learn-by-example macros, demonstrated a practical application of software that adapted to user behavior and simplified tasks.
Q: Is the AI hype cycle repeating itself?
A: Many parallels exist between the AI excitement of the 1980s and the current generative AI boom, including inflated expectations and concerns about misuse.
Did you understand? Charles Simonyi, the architect of “softer software,” also pioneered the pull-down menu, icons, and WYSIWYG editors – features we take for granted today.
Explore more articles on AI trends and future technologies. Share your thoughts on the evolution of AI in the comments below!
