• Temrel
  • Posts
  • Google scared of open-source AI?

Google scared of open-source AI?

Plus: Cool Tools, Machine Learning Lifecycle

May 08, 2023

Google leaked document suggests open-source fears

Luke Sernau, who has been a senior software engineer at Google since March 2019, wrote a message highly critical of Google's strategy. Based in Seattle, Sernau voiced concerns that Google's AI arms race with Microsoft could lead to its demise. Google, which had long been a pioneer in artificial intelligence, was challenged with the launch of ChatGPT by OpenAI.

“Open-source models are faster, more customizable, more private, and pound-for-pound more capable. They are doing things with $100 and 13B params that we struggle with at $10M and 540B.”

Sernau's message has been circulating within Google since April and has become a topic of conversation within Silicon Valley. He issued a warning that the open source communities are Google and OpenAI's true rivals, making faster progress and likely to beat the companies in the AI race. Sernau called on Google to open up to the open source community to avoid becoming obsolete and experiencing a strategic fiasco.

Cool Tools

Kadoa - helps crawl websites and extract data as JSON. I’ve done this before using just JS and it was just nasty. This tool makes it far easier. (link)

Midjourney - you’re probably already on this. Check out the image I created above. It’s an awesome product - once you get around the clunky Discord-based UX. Get it now while it’s only $120 a year. (link)

Machine Learning Lifecycle

These are the high-level strokes of the ML lifecycle. Pay attention - that ‘Refine’ feedback loop is what keeps the lights on.

  1. Problem definition: The first stage involves identifying the problem you want to solve and framing it in a way that is suitable for machine learning

  2. Data collection: In this stage, relevant data is collected from various sources and prepared for analysis. If you’re lucky, someone else is doing this step for you.

  3. Data preparation: This stage involves cleaning, transforming, and formatting the data to make it suitable for machine learning algorithms

  4. Data analysis: This stage involves exploring the data to identify patterns, trends, and relationships that may be useful for building a machine learning model.

  5. Model training: This is where the machine learning algorithm is applied to the data to create a model that can make predictions.

  6. Model evaluation: Once the model has been trained, it is evaluated to determine how well it performs on new data.

  7. Model deployment: The final stage involves deploying the model into a production environment so that it can be used to make predictions on new data.

  8. Model monitoring and maintenance: Once the model is in production, it needs to be monitored and maintained to ensure it continues to perform well over time.

Temrel is a devops consultancy. We help implement company’s infrastructure and operations to reach their ML goals. If this is something you need, we’d love to talk.