More Changes Inside OpenAI, Free Image Generation in ChatGPT, and Gemini for Brainstorming

PLUS HOT AI Tools & Tutorials

Aug 09, 2024

Hi!

So, last week's big question: healthy sleep or an interview of Lex Fridman with Elon Musk? (No, really, man, 8 hours...) If you haven't answered it yet, don't worry. In this post, we will highlight the most interesting statements from the podcast. And, of course, we'll discuss other news, like changes inside OpenAI and the beef in the AI community.

Let’s get started.

In this newsletter, we've assembled an AI Toolbox for Short Videos for you. Follow along if you want to simplify your content creation or try your hand at a new niche (without too many headaches):
AI Toolbox for Short Videos
Creators' AI
·
Aug 8
Read full story

This Creators’ AI Edition:

Featured Materials 🎟️
News of the week 🌍
Useful tools ⚒️
Weekly Guides 📕
AI Meme of the Week 🤡
AI Tweet of the Week 🐦
(Bonus) Materials 🎁

Featured Material 🎟️

More Changes (Or Problems?) Inside the OpenAI’s Office

The elimination game continues.

Earlier this week, John Schulman, one of the co-founders of OpenAI, left the company for rival AI startup Anthropic. The innovator who played a key role in the development of ChatGPT announced his decision at X. He added that his decision was motivated by a desire to deepen his focus on AI alignment—the science of ensuring that AI behaves as intended—and to do more hands-on technical work.

He later published another post with the following content:

I’ve decided to pursue this goal at Anthropic, where I believe I can gain new perspectives and do research alongside people deeply engaged with the topics I’m most interested in. I am confident that OpenAI and the teams I was part of will continue to thrive without me.

While the statement seems neutral, it's a pretty big stonewall against OpenAI and Sam Altman. Especially considering Schulman has moved on to a direct competitor.

The problem is that the news doesn't end there. OpenAI president and co-founder Greg Brockman has suddenly taken a break. The company has confirmed that he's taking a leave of absence until the end of the year. This is the first time in the nine years since launch. Brockman promises to return after he "relaxes and recharges."

Keep your mailbox updated with key knowledge & news from the AI industry

And That's Not All

Brockman, Schulman, and Deng. Who's next?

Schulman and Brockman are not the only ones who will not be working on OpenAI development in the near future. This week, The Information published an article titled "Trio of OpenAI Leaders Depart, Take Leave of Absence." You already know about the first two, but the third is product manager Peter Deng. Previously, he managed products at Meta, Uber, and Airtable.

OpenAI has already confirmed his departure.

Now add to that list the departures of co-founder Andrej Karpathy (founded Eureka Labs) in February, AI security researcher Jan Leike (now at Anthropic) in May, and co-founder Ilya Sutskever (founded Safe Superintelligence Inc.) in June. So, only three of the original 11 OpenAI founders remain: Altman, Brockman, and Wojciech Zaremba (head of language and code generation).

Should we be worried about OpenAI? I would say it's too early, but there are more than enough preconditions for that. The question here is whether Altman has managed to build a strong enough company that can operate effectively without some bright personalities. I guess only time will tell.
Perhaps OpenAI can recruit new visionaries to its ranks. For example, the company recently added Carnegie Mellon professor Zico Kolter to its board of directors. He is a big expert in AI security.

News Of The Week 🌍

DALL-E 3 Now Available for Free ChatGPT Users

Useful AI tools are becoming more and more accessible. Last year, OpenAI opened access to DALLE-3 functions for paid ChatGPT Plus subscribers. Now, image generation will be available to all users. With one (rather large) limitation: you can generate up to two images daily. One of DALL-E 3’s key improvements is that ChatGPT can develop a prompt to make an image, making it easier to make images.

As part of a Twitter announcement, OpenAI showed how ChatGPT helps a user discover what art nouveau looks like.

New YouTube Feature Lets You Use Gemini to Brainstorm Ideas

YouTube is testing integration with Google Gemini to help creators brainstorm ideas, titles, and thumbnails for videos. The Google-owned company announced the launch of its new Brainstorm with Gemini feature in a video published on the Creator Insider channel. A YouTube spokesperson said the feature is already available to some creators as part of a limited experiment.

The company wants to consider feedback from those users before launching.

Google isn't giving up on finding an AI product that will win over users. As TechCrunch notes, the new tool could give YouTube an advantage over other social video platforms often used by creators because it offers them something that those competing platforms don't.

Perplexity Reached 250 Million Queries Per Month

My favorite AI-powered search engine is breaking its own records. Perplexity answered about 250M queries last month, a huge increase from 500M queries for the year 2023. The company also revealed that it has achieved a monthly revenue rate seven times the amount it achieved in early 2024. Perplexity is on track to reach $35M in annual revenue this year, up from its previous expectation of just $5M.

And by the looks of it, Perplexity has no plans to stop there. The company recently announced a revenue-sharing model with publishers called the "Publisher Program" to expand the platform's reach. It has already signed agreements with Fortune, Time, Entrepreneur, Der Spiegel, and WordPress.

Reddit Teases AI-Powered Search Features

Reddit users will have access to AI-generated summaries. Steve Huffman recently told investors that the company is testing AI-powered search results pages to improve content recommendations. The company believes this will help users "dive deeper" into content and discover new Reddit communities. Reddit will use a combination of its own and third-party technologies to support the feature. The company will begin the experiment later this year.

All I can say is that it's long overdue. Reddit's interface is pretty outdated, and I personally have less and less desire to explore new subreddits.
Hopefully that will change soon.

Sharing is caring! Refer someone who recently started a learning journey in AI. Make them more productive and earn rewards!

Refer a friend

Serious Beef with AI Friends (Or Foes)

Nik Shevchenko vs Avi Schiffmann.

Over the past couple of months, a pretty serious (and yet very funny) conflict has been brewing in the AI community. Both young founders are developing AI-powered devices called Friend. Both started work in the spring of this year, and both accused each other of plagiarism. Since the conflict began, Shevchenko has managed to challenge his competitor to a fight and released a rap diss.

At the same time, Schiffmann managed to purchase the domain friend.com for $1.8 million (!) and called Shevchenko "an absolute piece of shit."

Both founders are developing a smart necklace, but the products have different philosophies. Schiffmann calls his device "a modern version of a Tamagotchi" that will help in the fight against loneliness. And Shevchenko focuses on a useful AI assistant that performs practical tasks. The device from the former creator is priced at $99, while the latter costs $49 for the developer kit and $69 for the consumer version.

Schiffmann's project is also clearly attracting investor interest. In July this year, the startup raised $2.5M. At the same time, Shevchenko wrote on his LinkedIn that he had raised investments twice, but did not disclose specific information.

I can't say I'm a big fan of conflicts between Influencers, but sometimes it's nice to dilute the news agenda with something emotional. If you want to understand the story a little better, Inc. has a good retelling of all the stages of the conflict between founders.

Highlights of Lex Fridman's podcast with Elon Musk

Especially for those who aren't ready to listen to an 8 hour 37 minute 34 second podcast, we've compiled some of the most important statements made by Elon Musk on the Lex Fridman podcast regarding Neuralink, AI, and other technologies.

Here's what we highlighted:

Neuralink implanted its device in a second patient paralyzed after a spinal cord injury while diving. By the end of the year, the company aims to have ten total human participants.
Musk said that within the next year or two, a person with a Neuralink implant will be able to outperform a pro gamer. The long-term goal of Neuralink is to improve human-AI interaction by increasing communication bandwidth and achieving speeds of up to 1 Mbps within five years, times faster than normal speech or typing.
Neuralink is already working on a second product. It is called Blindsight and should enable people who are completely blind or have lost both eyes or the optic nerve to see by directly triggering neurons in the visual cortex.
At a certain point, AI will be smarter than all humans combined. But even with super-intelligent AI, humans may still provide value as a "source of will."
Grok’s mission is to understand the universe. Grok 3 will likely be available by the end of this year.
It’s humbling to see how little data humans have actually been able to accumulate. Optimus robot is going to be the biggest source of data.
Musk believes that even seemingly harmless biases or inaccuracies embedded in AI systems could have disastrous consequences as AI becomes more powerful. He cited examples of chatbots providing incorrect or politically biased responses, highlighting the potential for these biases to be amplified at scale.

But if you have some free time, watch the whole thing. It's pretty interesting:

Useful Tools ⚒️

Inkeep – AI search & support powered by your content

Omnifact – Ultra-secure AI chat assistants for business

Kypso – Manage your team’s workflows with AI agents

Spreadsite – Turn Spreadsheets into interactive dashboards

Brainybear AI – Train AI Chatbots in 3 Clicks

If you need a good chatbot for consumer support but don't have the time (or desire) to deal with APIs, data analysis, and other complex tasks, you can try Brainybear. This tool analyzes content on your website or uploaded files and then uses them to advise users. One of the fastest ways to establish good feedback and customer service.