Welcome to the Tech Meme Right Home for Wednesday, July 24th,

2024. I'm Brian McCullough today.

Metas Lama 3.1 shows that open models can at least go toe to toe with proprietary AI models.

Why did the Google Wiz Deal fall apart?

Was the crowd strike outage a part of it?

Camilla Harris, AIsar, question mark, and what do the reviews of the recent Samsung

foldable phone say to us about those Apple foldable phone rumors?

Here's what you miss today in the world of tech. As alluded to yesterday, Meta debuted Lama 3.105B, the first frontier level open source AI model, according to Mark Zuckerberg, as well as new

Lama 3.170B and 8B models.

Meta also says it's working on Lama 4 already, which you'd expect, quoting

Ben's Bites.

Lama 3.105b is the biggest llama of them all.

Using it as a teacher meta has retrained

its 8 billion and 70 billion versions two together these models have a context

window of 128,000 tokens this means you can enter hundreds of pages worth of

content in your prompts. They're multilingual in eight languages, English,

German, French, Italian, Portuguese, Hindi, Spanish, and Thai.

They come with support for using tools like web search, math reasoning, and code execution.

They are open source.

You can download the weights and use them in your applications.

Lama 3.1 models are not multimodal, however.

They don't understand or create images,

...