DeepSeek's R1 Fundamentally Changes the Game - The Weekly #16

DeepSeek's R1 Fundamentally Changes the Game - The Weekly #16
"The low-hanging fruit is gone; future AI progress will require more complex technical breakthroughs."

Sundar Pichai, CEO of Google

As a subscriber to this newsletter, I'm pretty sure you'll have seen the news about Chinese firm DeepSeek and how they've trained their latest R1 reasoning model for an estimated $6m. This is far less than billions OpenAI has spent on their equivalent o1 model. But I'm not going to go into the details of that. In fact, it represents a good example of why I started Plain AI. The news that DeepSeeks' model performs better and cheaper than many other models out there is great and it should reduce costs for many parts of the AI sector, but for many of us, it's not entirely what matters.

Thinking about your use cases for AI and selecting the right tool for the job are fundamentally the most important things for you to consider. R1 is a reasoning model, meaning it's exceptionally good at solving mathematical and scientific tasks, similar to OpenAI's o1 model. But if you have a use case to predict customer churn, it's of little relevance. It also highlights that you need to build options into your workflows. With such a dynamic industry, it's important that you can swap out parts of your process as-and-when, without having to refactor the whole project.

In this week's newsletter, we cover:

  • Oracle gets into the Agents game
  • Italy is not keen on DeepSeeks AI model
  • Warnings that AI is a key driver for climate change

Let's get into it.


Curated News

Oracle Launches New AI Agents for Sales Teams

Oracle has just thrown its hat into the AI-powered sales game with a set of smart agents designed to take the busywork off sales professionals' plates. These AI helpers can handle the dull but necessary admin tasks—things like updating company records after meetings and pulling together customer intelligence reports from different business software platforms. Handy, right? Even better, they can work across multiple languages, making them a great fit for global businesses.

Oracle will be rolling these agents immediately and will not charge them extra. This move follows the trend from tech giants like Microsoft and Salesforce, who are all pushing AI agents to streamline workflows and boost productivity.

Italy Bans Chinese AI App DeepSeek Over Privacy Concerns

Italy’s privacy watchdog has just pulled the plug on DeepSeek, a Chinese AI app, over concerns about how it handles user data. Regulators weren’t happy with the lack of transparency around its data protection practices, so they’ve taken swift action—blocking the app and having it removed from Apple and Google app stores in Italy.

Experts Warn Of Extreme Artificial Intelligence Risks

A recent international report has highlighted significant concerns regarding advanced artificial intelligence (AI) systems. The study identifies potential issues such as job displacement, the facilitation of terrorism, and uncontrollable malfunctions. The International Scientific Report on the Safety of Advanced AI emphasizes the need for collaborative efforts among communities, governments, and companies to make informed decisions as AI technology progresses. The report categorizes the risks into malicious use, operational malfunctions, and widespread systemic issues, urging stakeholders to establish comprehensive frameworks to manage these challenges effectively.

Upcoming AI Events

AI & Big Data Expo 2025
London Olympia, Feb-5 & 6

AI UK 2025
QE2 Centre, London, Mar-17 & 18

Thanks for reading, and see you next Friday.

You can also follow me on social media: