Welcome back, my friends.

Grab your coffee — it’s Wednesday. This is the latest issue of /Imagine, your go-to source for the freshest news, tutorials, tools, and developments in AI. Let’s dive in…

🔮 Intel Announces Gaudi3 AI Chip

The Bytes:

✣ Gaudi3 is Intel’s latest AI chip announcement. With it, they’ve set their sights on NVIDIA.

✣ It’s projected to be 1.5x faster, have 50% more inference, and have a 40% efficiency boost vs. NVIDIA H100. They also claim it uses less power.

✣ Intel's strategic focus on AI has been years in the making, with Gaudi3 being their most advanced AI chip to date.

✣ They did not detail the pricing other than to say “We do expect it to be highly competitive” with NVIDIA’s latest offerings.

Why you should care: Intel's Gaudi3 chip is more than just another product release; it's a bold statement in the AI technology arms race, challenging the dominance of NVIDIA. This move not only showcases Intel’s commitment to AI but also heats up the competition, promising a new era of advanced, efficient AI computing solutions. This also comes on the heels of their $20 billion grant and loan infusion from the government. TLDR → Intel is making moves.

🌐 Google's Gemini Pro 1.5 Unleashed for Everyone on Vertex AI

The Bytes:

✣ Google’s Gemini Pro 1.5, the flagship of their GenAI model family, has entered public preview on Vertex AI for those with access.

✣ The model was launched in February, and we have covered it a ton here at /Imagine, but as a reminder, it has a 1 million token context length, supports 38 languages and regions, features enhanced reasoning and coding skills, and now can interpret uploaded audio and video streams.

✣ To access it, navigate to the Vertex AI section in your Google Cloud console. Look for “Model Garden” on the tools tab, select Gemini 1.5 Pro, and open in Vertex AI Studio.

Why you should care: Google's Gemini Pro 1.5 on Vertex AI represents a significant step forward in making advanced AI capabilities more accessible and pushing the boundaries of context length. If you have a chance, test the limits and give 1.5 Pro a try. It is pretty wild.

Oh, and Google dropped several other AI bombs at their Google Cloud Next 2024 Event, including Axion, their new ARM-based chip that claims a 30% better performance over competitors. Check out the full list at the TechCrunch summary above.

⚡️ Meta’s Llama 3: Gearing Up to Rival GPT-4

The Bytes:

✣ Meta is gearing up to release Llama 3, poised as a strong contender to OpenAI’s GPT-4, with a launch expected within the next month according to their president of global affairs, Nick Clegg.

✣ While specifics on the model’s size were not disclosed, Llama 3 is anticipated to boast around 140 billion parameters. Their goal is to make it the most useful assistant in the world.

✣ This replaces their lackluster Llama 2 which was released last year in July.

Why you should care: Llama 3's emergence signals Meta's ambitious stride into the generative AI race, aiming to match or surpass GPT-4's prowess. They have a long way to go, but rumors going around suggest this next offering is going to be pretty impressive.

OpenAI GPT-4 Turbo Update ⚡️

$2.9B Microsoft Investment 👀

Archetype AI Newton 🛠️

AI Health Assistant - Sequel 💪

New AI Tool Page Updates ✨

Prompt share: a person dressed like a Japanese demon. The person leans in and reaches toward the viewer as if to embrace. Hover. Angled view. A strange flapping and wavering portal is in the background --ar 4:54:5

Enjoy, my friends. Love you all. Have a great day!✌🏼

Join Us: Don't forget to follow me on Twitter/X at @jake_joseph and the brand on its fresh new page @slashimagineai. Let’s be friends.

Share The Insights: Do you know someone else who's obsessed with AI? Forward this newsletter and let's expand the community.

Want to advertise in /Imagine’s Daily Newsletter or on our Tools Page? → Book here!

Want more resources? →  Check out our Tools Page!

Feedback

Your perspectives help shape the content and hopefully help me make the newsletter suck less every day. Share your feedback on this issue or what you'd like to see next by directly replying to this email or by reacting to the poll above.