5 Things that Happened in AI this Week! 6.1.24

YouTube video (June 2, 2024) by Justin R. Lester ("that tech pastor") giving a five-item weekly recap of AI news: ChatGPT (ChatGPT4o advanced features rolling out to free users), rising LLMs (Gemini 1.5 Pro and advanced models climbing leaderboard), Elon Musk/X building toward a 10,000‑chip supercomputer to train Grok, Perplexity pursuing $250M at a $3B valuation, and practical prompting tips (six techniques). The host also promotes his website, AI Masterclass, and app-of-the-week (Meta AI).
Questions this video answers
June 2, 2024.
Raising $250 million at a $3 billion valuation (reported discussion).
ChatGPT4o (advanced features rolled out to free users).
Plans to build a 10,000‑chip supercomputer to train Grok.
Chapters
What's up everybody? It's that tech pastor here and we're here in the month of June and I want to do a recap of what's happened in the last week in tech and AI. Of course. Before we jump into this, don't forget to do a couple things. Number one, hit that like and subscribe button here at that. For me, Justin Lester, My hope is to give you everything in the tech world, at least summed up a little bit as I'm an AI adjacent leader learning as you're learning as well, to give you a tool to use and five things that are happening in the news when it comes to artificial intelligence. So please do me a favor and like and subscribe so we can have updates and hit that notification bell as well. If you haven't already ready, go ahead and make sure you go to my website, the link is in the show notes and I give you a list of my top 200 AI tools that you can use to enhance your business, enhance your work, enhance whatever you do to be all that you believe you can be for your job. Now let's jump into five things that are happening in AI this past week. First of all, Chat GPT unleashes advanced features. So the cool thing is Chat GPT4O is now available to everybody. That's YouTube, so it's rolled out advanced features for all this free user so you can browse the web, you can analyze data, you can upload your file, even explored visual content. Some of the tools I've talked about before like Pop Pro and even Claude, they were able to upload files, were able to do that on the free version of ChatGPT. Now you can. The crazy thing about this is that in spite of all of the growth of ChatGPT, Wikipedia has not lost any of its traffic. It's actually still resilient. People still want the knowledge from Wikipedia. It's interesting. Number two, Gemini 1.5 Pro and advanced models are climbing up the large language model leaderboard. The little chat bots that are coming out, Gemini 1.5 and advanced models are making waves and they're number two and three on the leaderboard. They're only behind ChatGPT4O users appreciate it and these models are improving daily. So if you haven't tried a Gemini based or anthropic based chatbot, go check it out. Number three, Elon Musk. Now we've GROK has been only available to those who paid on X. So Elon Musk is never one to shy away from ambitious Products now projects and is added again. Xai Tim wants to build a 10,000 chip supercomputer. The goal is to train Grok AI model. Imagine the power behind not only elon Musk and AI but 10,000 chips supercomputer. It's going to be interesting. I'm curious to see what happens. Be on the lookout and be understanding and looking and researching as this is happening. Lastly, Perplexity. Now you guys have heard of my love and care for Perplexity. If you've been on this channel at any time. I love what Perplexity does and how it's able to do amazing research for you. Here's the thing about Perplexity. Perplexity is a rising star on the LLM space. They're in discussions and to raise another $250 million at a 3 billion dollar valuation, investors are betting big on Perplexity. Once you bet big too. If you haven't used Perplexity, go check it out. And finally number five and the link is in the show notes too. Prompting Tips so the thing about AI is it's all about the data. You put in. You put in bad data, you get bad responses, you put in good data, you get good responses. Here in your prompting tips I want you to see there are six different techniques that are out that some researchers got together and these are the best ways to do so. Number one in shot prompts and in context learnings they suggest providing five examples to better prime the model. Number two chain of thought prompting to push the model to explain its reasoning. Number three Retrieval augmented generation to allow the model to reference relevant text for improve accuracy. I use that format a lot. Number four structured input and outputs using markdown context for clear context. Number five Specific and multi step instructions to give the model specific and detailed prompts and rules to follow. And essence of literally telling it respond like this, talk to me like this, engage with me like this, things like that. Really having details in the ways that you give prompts. And then lastly break down complex tabs. So asking it. This is the big piece. Let's start with step one and then break down step one. Start with step two step break down step two. The reason being remember AI is all about the data you put in. You put in bad data, you get bad responses, you put in good data, you get responses. Your prompts are what are going to tell it. So be intentional with your prompts and if you take my masterclass I save a prompt worksheet that I'll give you in the Masterclass. Go to JR Lester.com click on AI you see the AI Masterclass and in that masterclass I give you a list of prompts on how you yourself can improve how you prompt. We do practices together and I give you a free console. We have a chance to actually have a conversation about how to use it in your business because we're giving you the business for your business. So what is the app of the week? If you haven't tried Meta, you guys have heard my love for Copilot. If you haven't tried Meta's AI Meta AI is awesome. It's an open source LLM instead of close. So like open source. In that case, it's, it's continuing to evolve, it's continuing to grow and it's based upon llama 3. It's so, so powerful. It's encouraging. Opening ended conversation, answer queries and you can even attempt coding tasks. So give it a try. You can create images, you can create code, you can create a lot of things on meta. On meta AI. So I love it. I actually have it booked on my computer and so I use that along with Copilot to play around with. Try meta. Don't forget like and subscribe. Come back each and every week for an update on what's going on in the tech world. I'm an AI adjacent leader. I'm not a tech professional professionally. I pastor a church. I just love this stuff. I've been in it for years and I'm just two feet all the way in an artificial intelligence. And as I'm learning, I want to bring you along on the journey. So the app I suggest you try, go try Meta if you haven't already. Follow me on Instagram and on Facebook at thattechpastor. And we have daily updates, sometimes two, three times a day on everything that's going on. A tool to try. Some are good, some are bad, but I try it so you don't have to and I try it so you can have fun when you do it. All right, I'll see you next week for another update in the top five things that happen to AI this week. God bless you.
More from Justin

AI gives you hours back every week. Most people's first move is to fill those hours with more work.

Most people stall out on AI at the exact same step. They don't know where to start.

Thank you for the follow-ups, the messages, the emails, all of it since the New York Times piece. Gr

The New York Times spent a weekend with me. Digital twins, AI, the local church, all of it.






