GM Readers! ☀️

Welcome to the 11th issue of Evolving Internet Insights — a weekly newsletter curating the top stories and our insights on the “Evolving Internet,” which covers AI, Web3 and everything in between.

What we are watching: there are articles floating around like this one that say Microsoft is interested in acquiring Nintendo. 🎮

Thanks for reading!

Liang and Dan 🙌

🌰 In a Nutshell

• Google is building its most advanced LLM model to dethrone OpenAI’s GPT4

• Open AI announces their next-gen image model, DALL-E 3 and will integrate it directly into ChatGPT

• 98% of CEOs think AI is impactful, but worry about potential erroneous results

• More and more presentations are made with AI

💾 Byte-sized Stories

This week’s top stories with our insights on top.

1. Google Is Building Its Most Advanced LLM 🤓

⚡️ TL;DR: Google is around the corner from launching Gemini – its most advanced LLM – to compete with OpenAI. Gemini is going to be multimodal which means it can generate and integrate text, images, and other data types. Gemini is being developed by the Google Brain Team and Deep Mind (read: the elite of the elite at Google).

⚡️ So What: Demis Hassabis, the CEO of DeepMind, has mentioned that Gemini will eventually be able to utilize memory and search history and can fact check against Google search to minimize a model’s “hallucinations.” The long term vision is for Gemini to be a “universal personal assistant,” that can help you with a little bit of everything–a tool that would make the utility of current chatbots feel trivial.

⚡️ Zoom Out: Every big tech company is looking to dethrone OpenAI and GPT4. While OpenAI seems to have a first mover advantage, companies like Google and Meta might be able to make a comeback and generate a second mover advantage that comes from seeing all the flaws and issues with GPT4 (Apple has proven time and time again there is a such a thing as a second mover advantage).

Read More Here, Here

2. OpenAI Launches DALL-E 3 and Integrates It Into ChatGPT ⚔️

⚡️ TL;DR: OpenAI announced their next generation image model, DALL-E 3. DALL-E 3 will understand more nuanced user inputs to produce images closer to what is being prompted. OpenAI will integrate DALL-E 3 directly into the ChatGPT interface and it will be available in October to ChatGPT Plus and Enterprise users.

⚡️ So What: By integrating DALL-E 3 directly into ChatGPT, OpenAI is hoping to make ChatGPT the go-to interface for everything. Just like Google Search became the de facto place most people start their workflow to navigate the Internet. Many companies are now focused on stitching together different models together in one interface. Our prediction is that the next big “wave” in generative AI will be building multimodal interfaces.

⚡️ Zoom Out: OpenAI also launched a commercial that gave us Google “Year in Search” video vibes. OpenAI is trying to simplify AI for the average user, and, in the process, make it more approachable. This is a smart move because the average user doesn’t care about the tech, they care about how it can help them. Building game changing technology is just one part of the equation, mass adoption is a whole other thing.

Read More Here, Here

3. 98% of CEOs Say Implementing AI is Beneficial, But Worry About Potential Errors ⛔

⚡️ TL;DR: Workday, the workflows giant, conducted a study of CEOs and C-suite executives to understand how they view AI’s impact on their organizations. Here are some of the key findings from the report:

• 98% of CEOs said there would be some immediate business benefit from implementing AI.

• 47% of all business leaders believe AI and ML will significantly amplify human potential.

• 43% of all business leaders are concerned about the trustworthiness of AI and ML.

• 59% of respondents said their organizations' data is somewhat or completely siloed, and only 4% of all respondents said their data is fully accessible.

⚡️ So What: We have posed the question a few times: what is preventing businesses from adopting AI? While it seems like every company sees promise in AI, there are still some problems:

• Organizations aren’t structured to implement AI

• Leaders are concerned about AI potentially making errors that could be harmful in a variety of ways (who would they “point the finger at” if something goes wrong)

• Organizations that don’t have good data housekeeping processes (easily accessible, not siloed, structured correctly, etc) will struggle to implement AI. As we used to say in consulting, “garbage in, garbage out.”

⚡️ Zoom Out: While the AI funding cycle for startups shows no signs of slowing down, it is important to understand how adoption works at larger companies (read: an AI startup’s customers). Large organizations have a lot more to lose than gain when adopting at the onset of an emerging technology cycle. Startups need to appreciate that beyond the technology they are building, sales and business development are just as important.

Read More Here, Here

📊 Let’s Get Graphic

One visual we couldn’t stop thinking about.

Search Interest For The Term: “AI Presentation Maker”

⚡️ Takeaway: For better or for worse, we make a lot of presentations for work. With the rise of generative AI tools, now professionals can use AI to help them automate tedious and boring tasks at work. There will be a few interesting things to watch:

• Are AI generated presentations higher quality than human created presentations?

• Will employees get more productive or will their bosses just ask for more presentations once they realize employees can churn out presentations at increasing speeds?

🐇 Down the Rabbit Hole

Some deeper dives to help you get smarter on emerging tech.

1. Open Source Software Stats: a good list of stats that highlights how much of the Internet runs on open source software.

2. Guide to LLM Training: a (technical) guide by Replit outlining how you can train your own LLM (large language model).

3. Economic Case for Generative AI: a16z deep dives into how generative AI and its capabilities will usher in a new economic paradigm.

🗳 We’d Love your Feedback

🙏 Shameless Asks

It takes us days to put this together, sharing it takes you 19.45 seconds 😘

An easy way to support us and the newsletter is through the referral program.

⚡️ Share Evolving Internet Insights with a friend and ask them to subscribe

⚡️ Share on Twitter and Linkedin with a short note

⚡️ Share on your company Slack/Teams channels and communities