AI summer is over
What will AI fall have to offer us?
There are a number of indicators that the AI market is cooling. The recent fall in Nvidia stock and profits combined with commentary on bloated valuations of the magnificent seven seem like leading indicators for a shift in investor attitudes towards AI products. This is mirrored in public sentiment. The lackluster release of GPT-5 lead business insider to dub today’s tech climate as the “AI meh era.” There is growing public skepticism about the safety and effectiveness of AI tools on the market, demonstrated by articles about AI psychosis and AI related deaths, and children being exposed inappropriate content. Throw in the MIT report that found that 95% of AI pilot programs fail, and I’d say it’s safe to assume that AI summer is coming to a close.
Like any good New Englander, I feel the end of summer is not a thing to grieve. In fact, I look forward to fall in every way… Including the ushering in of AI fall. I think as the hype and FOMO fades a bit we will start to see the true value of AI tools and what goes into making them safe, effective, ethical, and impactful.
AI fall
Here are a few things I think we will start to see as the market slows down:
Companies will reinvest in humans: this is something we are already seeing, and in some cases it looks downright hypocritical since the companies touting the use of AI to replace employees are now bringing on more employees than ever. It won’t be the same number of employees, but hiring will increase in product management, design, and research.
Go to market will be more refined: with that reinvestment and a decrease in FOMO, the tools going to market will have an improved UX and will come out more fully formed. We will start to see an increase in successes post launch. My hope is that this doesn’t create a FOMO resurgence, but instead reinforces, good behavior and practice. Companies putting out good case studies on how they found success will help fuel this.
The kinds of tools will diversify: so many of the startups out there that haven’t gone to market are drilling down on more specific problems rather than focusing on time saving and efficiency. We will see an increase in tools that are more effectively solving real world problems for more specific types of users.
Data will be proprietary again: fewer companies will be going to market as a wrapper over an API call to OpenAI. They will own their data and models. This may cause some hiccups immediately after release as these models may need an increase in data to become more effective. Companies that take the time to use beta testing to develop a larger data source and refine outputs will be more effective at go to market.
Ethics and governance will be the new data privacy: just like companies had to dial in privacy statements and practices, they will be increased pressure for AI tools to launch with transparency statements, governance, and a focus on gaining user trust. Users will be more savvy in looking for this and expecting it.
The net result of all of this will be better products on the market, providing higher value. The winners will be companies that took more time getting to market, including smaller startups that were more focused and took greater care to release more high-quality products, and larger companies adding AI to their overall offerings. The real winners, however, will ultimately be the users.
