Keeping Up With Data #117
5 minutes for 5 hours’ worth of reading
The economic outlook is not overly optimistic. I sat at a conference in Phoenix on Wednesday where an economist from ITR Economics was predicting a mild recession for 2023.
The times of recessions offer companies a unique opportunity to modernise, digitalise, and automate. Opportunity to make improvements they often don’t have capacity to focus on when their business is flying.
It is interesting to see the chart from Bain’s Private Equity Report confirming that “deals done through a downturn generate superior returns over time.” Private equity isn’t shy of making investments to grow their companies.
For the investors it means that “waiting risks losing a valuable opportunity to profit from the rebound.” For companies waiting risks losing a valuable opportunity to improve.
As a data professional, I reserve the right to remind everyone of the great value-creating potential of data and analytics. Now is the time to invest into your data capabilities and navigate the recession better than competition and ready to fly again soon.
Today’s reading list looks at ChatGPT and its move along the hype cycle.
- The inside story of how ChatGPT was built from the people who made it: The success of ChatGPT came as big surprise to its creators. The technology wasn’t new and the research preview going viral wasn’t really expected. The global success resulted in dozens of millions of people playing with the technology providing useful feedback into what works and what doesn’t. Being at the forefront often comes with a necessity to cut corners. Indeed, one of the co-founders admitted the company “underestimated the extent to which people would probe and care about the politics of ChatGPT.” (MIT Technology Review)
- Addressing criticism, OpenAI will no longer use customer data to train its models by default: “Starting 1st of March, OpenAI says that it won’t use any data submitted through its API for ‘service improvements,’ including AI model training, unless a customer or organization opts in.” The company is also “simplifying its terms and data ownership to make it clear that users own the input and output of the models.” This is a big step in a direction to make the technology friendlier to developers who were complaining about the privacy risks involved. It of course limits the ability to improve the machine learning models, but it serves the purpose of becoming a pick-and-shovel play. (Tech Crunch)
- Generative AI: A Creative New World: A piece from September 2022 about generative AI. Towards the last year, we saw some investors shifting from web3 to generative AI on the back of the success of Dall-E 2, Stable Diffusion, or Midjourney. This trend has obviously been magnified by the success of ChatGPT. We are somewhere behind the peak of inflated expectations phase of the technology hype cycle. And now people start to see the downsides and some have a creeping feeling of a déjà-vu of IBM announcing all problems soon to be solved by their Watson. The technology is here and it’s up to companies to develop better models and their applications now. (Sequoia)
Enjoy the weekend and remember that keeping up with data is easier than catching up.
