Are we running around with the generative AI hammer, pretending everything is a nail? Are AI startups outpacing corporations while prominent tech moguls announce new features months or years from becoming real while their teams catch up and release real solutions?
We hear a variation of these two questions almost daily during our executive briefings and workshops, so let’s unpack this.
Entrepreneurs are paving the way by showing what is possible:
Generative AI is relatively simple to master and easy to value. Consequently, usage has grown, leading to practical solutions like chatGPT among the 10,000+ tools we’ve mapped at AI4SP.org.
In our analysis, most of these AI solutions, for example, Perplexity AI, SlidesAI, Type.ai, or even an enterprise-class platform to address privacy and compliance when using AI such as 2021.ai, were created by nimble startups, outpacing large companies.
This phenomenon places additional pressure on large companies that scrambled to either release half-baked solutions or a ton of PR with promises of what was coming.
The GPT Hammer:
It is still the early days, and things are confusing, to say the least. We’ve got an army of consultants, entrepreneurs, and independent software developers wielding the GPT hammer, seeing every problem as a nail. But let’s be clear: GPT isn’t always the answer.
Sometimes, a model that’s 50x cheaper, like Ada, with the proper embedding and fine-tuning, is the perfect solution for use cases requiring specific knowledge with minimal hallucinations or confabulations.
Through our contact form at AI4SP.org, we receive a dozen AI startup pitches every week at AI4SP.org. Many start with “Imagine GPT for x,” and a percentage have fundamental flaws in their assumptions because they are running with hammers in search of nails.
A PR Game from Large Corporations
A recent post from Gartner’s executive Rajesh Kandaswamy reminds us that selling vaporware isn’t new in tech. However, due to the volume and variety of impacted products, the generative AI gold rush has amplified this practice.
As an executive at ABB, Microsoft, and Yahoo, and advising Venture Capital firms, I was right in the middle of the early days of the PC revolution and the .com boom. The PC and Internet industries were well-known for slowing down their competitors’ sales by announcing better features before they were ready.
But then, software development cycles were rather long, and we did not enjoy the immediate gratification of online solutions and mobile apps.
Today, agile entrepreneurs learn by execution. Since there is no legacy user base to care for, they quickly release AI-powered solutions that deliver on the large company’s promises. Two examples:
- Perplexity AI Copilot, which is our favorite Generative AI assistant so far.
- Canva, with over 100M global users, is a core productivity tool for companies of all sizes, and they have been incorporating AI features rapidly. Many features were acquired from nimble entrepreneurs.
It’s An Enterprise-enabled and Small-Business-led race:
While this is an enterprise-enabled revolution (Msft, Google, AWS, OpenAI, Alibaba, IBM, NVIDIA), it is an SMB-led revolution enabled by these giants. SB innovators move faster and are closer to customers’ wants and needs.
The old models we invented in the software industry, including the pre-announcement tactics or monetization models, need some rethinking.
Some leaders at Fortune 500 companies that were masters at the license-selling and subscription models games or that still think of long development cycles to release 1,000 features packed in a rich solution can borrow a page from the SMB AI-creators playbook.
And we all need to realize that GPT is not a hammer in search of nails, as the potential for AI to drive social progress, equity, and economic growth is vast. Let’s be sure we use the correct set of tools responsibly to create, use, and support AI that works for all.