#14 Can you scrape the internet for free?
Reddit & Stack Overflow to charge companies for using their data to train big AI models, and 35 ways in which real people are using AI right now.
If you are new to Discovery Unlocked, here is our one-line pitch-
We talk about the most exciting 🛠️ tools/launches, 🐣 thoughts and 📑 developments of the week, on Friday. Only one email per week, so no “AI fatigue”.
Table of contents
🛠️ Tools
Lesson Plan Generator [link]
Clear- AI Journal to measure and elevate your well-being. [link]
Banter AI- Human-like conversations with any AI celebrity [link]
Bloks- AI-powered productivity assistant [link]
Phind- AI search engine for developers [link]
Hushl- AI assistant to help you write creative and insightful Linkedin content [link]
Debate Devil- Challenge your hypothesis with devil’s advocate arguments [link]
📑 Developments
Stability AI, makers of Stable Diffusion, launch StableLM suite of language models [link]
DeepMind and the Brain team from Google Research will be joining forces as a single, focused unit called Google DeepMind. [link]
Adobe Lightroom introduces AI innovations to make edition photos easier [link]
Weblfow AI-design and build websites using natural language [link]
Atlassian Intelligence, an AI-powered virtual teammate.[link]
Humane, a $230 million funded startup, previews AI-powered wearables at TED [link]
Adobe brings Firefly to its video tools [link]
Alphabet shares fall 4% after Samsung considers replacing Google with Bing as its default search engine. [link]
Reddit & Stack Overflow want to get paid for helping to teach big AI systems [link] [link]
Accounting automation startup Trullion lands $15M investment [link]
+ 🐣 Thoughts
Complete beginners guide to autonomous agents, Moving forward size of LLMs won’t matter as much, TED talk- the inside story of ChatGPT's astonishing potential and more.
+ 🥷 Random
35 ways in which real people are using AI right now, an example of LLM prompting for programming, artist rejects photo prize after AI-generated image wins award and Midjourney AI guide.
🛠️ Tools
Lesson Plan Generator [link]
Clear- AI Journal to measure and elevate your well-being. [link]
Banter AI- Human-like conversations with any AI celebrity [link]
Bloks- AI-powered productivity assistant [link]
Phind- AI search engine for developers [link]
Hushl- AI assistant to help you write creative and insightful Linkedin content [link]
Debate Devil- Challenge your hypothesis with devil’s advocate arguments [link]
🐣 Thoughts
"We’re democratizing Generative AI so companies of all sizes can leverage it.”, Andy Jassy, CEO of Amazon [hot take]
Read Andy Jassy’s Letter to Shareholders [link]
“Size of LLMs won’t matter as much moving forward“ & “There's no GPT-5 to worry about...yet”, Sam Altman, CEO of OpenAI [hot take]
Sam Altman in his talk at MIT said that he sees size as a false measurement of model quality and compares it to the chip speed races we used to see.
“I think there’s been way too much focus on parameter count, maybe parameter count will trend up for sure. But this reminds me a lot of the gigahertz race in chips in the 1990s and 2000s, where everybody was trying to point to a big number”
He also has squashed rumors that OpenAI is already working on GPT-5.
Read here [link]
Complete beginners guide to autonomous agents [Blog]
Read here [link]
OpenAI co-founder Greg Brockman on “The inside story of ChatGPT's astonishing potential” [TED talk]
Brockman explores the underlying design principles of ChatGPT and demos some mind-blowing, unreleased plug-ins for the chatbot that sent shockwaves across the world. After the talk, head of TED Chris Anderson joins Brockman to dig into the timeline of ChatGPT's development and get Brockman's take on the risks, raised by many in the tech industry and beyond, of releasing such a powerful tool into the world.
Watch the talk here [link]
Why AI will not steal our jobs [Hot take]
Read it here [link]
📑 Developments
Stability AI launches StableLM suite of language models [launch]
StableLM is trained on a new experimental dataset built on The Pile, but three times larger with 1.5 trillion tokens of content. Details on the dataset will be released in due course. The richness of this dataset gives StableLM surprisingly high performance in conversational and coding tasks, despite its small size of 3 to 7 billion parameters (by comparison, GPT-3 has 175 billion parameters).
Read more here [link]
Google DeepMind [restructuring]
DeepMind and the Brain team from Google Research will be joining forces as a single, focused unit called Google DeepMind. [link]
Adobe is unveiling new AI innovations in the Lightroom [launch]
New AI innovations in the Lightroom ecosystem make it easy to edit photos like a pro
Read more here [link]Weblfow AI-design and build websites using natural language [launch]
New AI capabilities within the Webflow Designer will immediately offer personalized assistance whenever you need it. Users will be able to harness the power of Webflow University and other educational content by simply asking for help as they go to get advice and support with responses based on their specific needs.
Read more here [link]
Atlassian Intelligence [launch]
It can summarize decisions & action items, define test plans for products, and do other forms of text retrieval and generation. Read more [link]
Humane’s co-founder gave an early glimpse of the AI-powered wearables [demo]
In his TED talk, Chaudhri showed the wearable, which sat in his jacket pocket, translating his own voice into French. He also answered a phone call from his wife with the call information appearing as a green image projected onto his hand.Read here [link]
Adobe brings Firefly to its video tools [link]
Samsung Electronics was considering replacing Google's search engine [news]
Google parent company Alphabet Inc. shares fell nearly 4 percent on Monday following a report over the weekend that Samsung Electronics was considering replacing Google's search engine with Microsoft's Bing on its devices. [link]
Reddit & Stack Overflow want to get paid for helping to teach big AI systems [news]
OpenAI, Google, and other companies building large-scale AI projects have traditionally paid nothing for much of their training data, scraping it from the web. But Reddit & Stack Overflow, both popular internet forums, plan to begin charging large AI developers for access to the content on their platforms.
Accounting automation startup Trullion lands $15M investment [fundraise]
Trullion is an AI-powered platform that automates manual work for finance and audit teams—ensuring accuracy, compliance, and total confidence. It is based in New York and was founded in 2020. The company has raised $33.5 million to date.
Read more here [link]
🥷 Random
From planning workouts to designing parts for a spaceship. 35 ways in which real people are using AI right now [use cases]
Read here [link]
Viral AI-powered Drake and The Weeknd song is removed from streaming services [news]
An AI-voiced rap song replicating the voices of the artists Drake and The Weeknd that went viral on Sunday had been taken down from streaming services as of Tuesday morning.
The song, “heart on my sleeve,” was created by an anonymous musician who goes by ghostwriter. “Heart on my sleeve” features vocals that sound like Drake and The Weeknd as well as Metro Boomin’s signature producer tag. Clips of the song went viral on social media, with one Twitter upload of the song garnering 6.9 million views.
Read more [here]
Artist rejects photo prize after AI-generated image wins award [news]
German artist, Boris Eldagsen, has rejected an award from a prestigious international photography competition after revealing that his submission was generated by Artificial Intelligence (AI).
Read here [link]
Reverse psychology to trick ChatGPT
People are using AI for therapy, even though ChatGPT wasn’t built for it. [news]
The idea of using a chatbot in a therapeutic or coachlike manner isn’t without precedent. In fact, one of the earliest chatbots, Eliza, was built in the 1960s by Joseph Weizenbaum, a professor at the Massachusetts Institute of Technology, to imitate a therapist. Several chatbots, such as Woebot and Wysa, now focus on mental health. Unlike human therapists, chatbots never get tired, and they’re inexpensive by comparison.
But there are also potential pitfalls. Powerful general-use chatbots including ChatGPT, Google’s Bard, and Microsoft Corp.’s OpenAI-powered Bing Chat are based on large language models, a technology with a well-documented tendency to simply fabricate convincing-sounding information. General-use chatbots aren’t designed for therapy and haven’t been programmed to conform to the ethical and legal guidelines human therapists observe. In their current form, they also have no way to keep track of what users tell them from session to session—a shortcoming that most patients probably wouldn’t tolerate from their human therapists.Read more here [link]
An example of LLM prompting for programming [prompting]
Xu Hao uses a chain of thought and general knowledge prompting with ChatGPT when writing self-testing code.
Read here [link]Midjourney AI Guide [resource]
Find the guide here [link]
We hope you liked what we have featured. As always we cannot stress more how eager we are to hear from you. Your remarks, insights, and feedback would be super helpful, so please tell us what you liked, or what you didn’t.
Please, do tell your friends about us.
Here’s wishing you all a happy weekend.
See you again, next week.
Very informative!