AI News of the week – 21st April – a quick summary

The big take away from last week was that Incumbents cannot be counted out. Whether it’s Meta (and yes, they are an incumbent at this point) or Adobe, they’ve both stayed relevant to the AI conversation through everything. We’re particularly proud of Meta’s contribution because its open-source. We also talk about updates from Google, Boston Dynamics and the world of AI wearables.

Let’s jump into the major updates from last week:

• Meta launched Llama 3 – their open source LLM that builds on the their existing Llama 2. Llama 3 – 8B is the best small model out there, currently beating out the other small models. Their 70B model is also highly competitive. We no longer pay attention to the jump in improvements because this is going to be a race with updates every few weeks / months that will improve on various benchmarks. More important for us is the fact that there are those pushing this via open source.
  
  You can read more about Llama 3 here. You can also try it in action via Meta AI. Bonus: Here’s an amazing interview with Mark Zuckerberg. We always enjoy him being candid.

• Meta AI is coming to WhatsApp and all Meta products are going to get a massive upgrade. You can chat with Meta’s AI within WhatsApp. It’s currently being rolled out in a phased manner. We had a chance to test it and it’s definitely better than our experience with ChatGPT 3.5. We’re guessing this is because of Llama 3. More on Meta AI’s update explained below:

• Adobe announced updates for Premiere Pro powered by the Adobe Firefly video model. You will soon be able to add or remove objects, extend parts of your canvas, etc… You can now do what Sora, Runway ML and Pika Labs do for you, all within the Premier Pro workflow itself. This is an absolutely massive move that will make life for those who are comfortable in their existing Adobe work-flows much easier.

• Limitless AI launched a wearable AI that might fill the void that Humane’s AI pin left. They reached over 10k orders within 2 days of launching. No small feat given the overall sentiment on wearable AI. Goes to show that the comments about Humane’s AI pin were not personal and there’s a very big market for AI wearables.

Here’s the Co-founder, Dan Sirokar introducing it:

• Boston Dynamics announces an all-new Atlas robot. As pioneers in the world of robotics whose videos have caused equal amounts of excitement as they have nightmares, Boston Dynamics is now ready to push the boundaries even more.
  
  Meet the new Atlas robot. The video is a reminder in the fact that a Humanoid Robot, despite being as human-like as it is, doesn’t have to be exactly like humans. With the pace of progress in the world of humanoid robots, we are definitely getting closer to a Jetsons-like future than ever before. Either that or the Omnic Crisis.

• Google Published a memo from Sundar Pichai on the building for the future. The whole blog post can be summarised as:

• Google is restructuring its AI teams to improve collaboration and speed up innovation.
• AI research teams are being consolidated into Google DeepMind.
• A new Platforms & Devices group is being created.
• The goal is to make Google a leader in safe and responsible AI.
• Google is committed to developing AI that is beneficial to society.

That’s a wrap on this week’s posts. If you like this, please consider sharing this with others who will appreciate it. We’d also love to hear back from you on what else you’d like us to cover. Check us out on LinkedIn and drop us a comment there.