Pre-training has continued to evolve with the emergence of foundation models such as BERT, GPT, DALL-E, CLIP, and others. These models are pre-trained on large general-purpose datasets (often in the order of billions of training examples) and are being released as open source by well-funded AI labs such as the ones at Google, Microsoft, and OpenAI. 

They allow startups, researchers, and others to quickly get up to speed on the latest machine learning approaches without having to spend the time and resources needed to train these models from scratch.

One of the risks associated with foundation models is their ever-increasing scale. Neural networks such as Google’s T5-11b (open sourced in 2019) already require a cluster of expensive GPUs simply to load and make predictions. Fine-tuning these systems requires even more resources.

More recent models created in 2021-2022 by Google/Microsoft/OpenAI are often so large that these companies are not releasing them as open source – they now require tens of millions of dollars to create and are increasingly viewed as significant IP investments even for these large companies.

Dataset alignment can also be a challenge for those using foundation models. Pre-training on a large general-purpose dataset is no guarantee that the network will be able to perform a new task on proprietary data. The network may be so lacking in context or biased based on its pre-training, that even fine-tuning may not readily resolve the issue. 

Any startup leveraging foundation models in its machine learning efforts should pay close attention to these types of issues.

The opportunities and risk around using hosted and pre-trained models has led many companies to leverage cloud APIs in the “experimentation phase” to kickstart product development.

Once a company has determined it has a product-market fit, it often transitions to self-hosted or self-trained models in order to gain more control over data, process, and intellectual property. This transition can be difficult, as the company needs to be able to scale its infrastructure to match the demands of the model, as well as manage the costs associated with data collection, annotation, and storage.