- AI In The Middle
- Posts
- They’re coming for you Sammy boy
They’re coming for you Sammy boy
Baby AGI spotted
Happy Friday! It's International Moment of Laughter Day. Your homework for today is to laugh out load at least once. 😀
Everyone has of course been talking about autonomous AIs. I covered Auto-GPT and Baby AGI last week, so won’t dive too much into the aftermath this week. Suffice to say, it only took someone a few days to create a GPT on a mission to destroy humanity.
Somehow, destroying humanity came higher on the list of use cases AI can take on compared to solving the UN Sustainable Development goals. There is a lot to worry about and I’m leaving it to Seth Herd to give you the bad news, as I’d like to focus on happier develpments.
Now that we’re aligned (you’re a nerd if you laughed), here’s what I have for you today:
🤖 Poe, a Quora product, now allows users to create their own chatbot by combining prompts with existing bots like ChatGPT, making it a competitor to Character.AI. Chatbots with "personality" are on the rise.
🧑🤝🧑 Google and Stanford researchers have created an AI-driven world called Smallville, where 25 AI characters live and interact with each other, showing that AI can make game characters more engaging and believable.
👓 Stanford students created a tool that uses AI and AR to generate responses during conversations. It can help with social anxiety, public speaking, and job interviews by displaying the generated response as if it came from the individual wearing AR glasses.
User-created bots: “personality” drives retention
A few days ago, Poe announced that users can now create their own chatbot by combining prompts with an existing bot such as ChatGPT. Poe is a product of Quora, which launched in February of this year. According to the CEO of Quora, Adam D'Angelo, the future will bring a proliferation of bots, not just general ones like ChatGPT or Claude, but more specific bots as well. In fact, people have already created bots for trip planning, learning math, and even flirting on the platform.
This move by Poe makes it a direct competitor to Character.AI, another AI startup that generates chatbots capable of mimicking different personalities, from celebrities to fictional characters.
It's interesting to see how chatbots with a "personality" are becoming increasingly popular. They're more engaging compared to bland bots like chatGPT (sorry Sammy boy). Looking ahead, it's possible that we'll get to interact with bots with more advanced personalities, with Gordon Ramsey yelling at us in the kitchen, Master Yoda guiding us through tough decisions, or David Goggins motivating us to start the day with 10 push-ups.
AIs turn out to be pretty good at being human
So, Google and Stanford researchers whipped up this AI-driven world called Smallville, where 25 AI characters live and do their thing. They've got daily routines, can remember stuff, and some even got into politics or picked up careers on their own. This whole thing could make video game characters way more interesting and engaging.
Smallville looks like an old-school video game, and the characters chat, make friends, and even throw parties. Researchers think this project shows that AI can help make characters in games and stories way more believable and fun.
Of course, there are some concerns with bringing this kind of thing to the real world. People might get too attached to these AI characters, or mistakes could lead to wrong conclusions about what users want. Plus, there's always the risk of misinformation or harmful content being created.
Sounds like this could just give the metaverse a second life (see what I did there)?
AI-Enhanced Conversations with chatGPT & AR Glasses
meet lifeOS: an operating system for your entire life 🌐
a personal AI agent delivered directly through AR smart glasses 👓
it uses computer vision to 👁️recognize👁️ your friends’ face
then brings up relevant details to talk about based on your texts with them (memory🤯)
— Bryan Hau-Ping Chiang (@bryanhpchiang)
6:57 PM • Apr 10, 2023
A group of Stanford University students have developed a tool that utilizes AI and augmented reality to assist individuals during difficult conversations. This tool leverages GPT-4 language model and Whisper, an OpenAI voice recognition tool, to generate responses to questions asked during the conversation. The generated response can be displayed by wearing AR glasses with a monocle as if the individual came up with it themselves.
The team has named this technology ‘real-time Charisma as a Service (CaaS)’ and believes it has many potential applications, including helping people with social anxiety, public speaking, and job interviews. Whisper's AI-based speech recognition is utilized to listen to conversations and generate appropriate responses based on the context.
It all looks very janky and not many are ready to wear glasses with monocles, but the project is enabling us to start imagining what the future will bring.
Gen AI Deals that make your eyes (and mouth) water 💰
What caught our eye? 👀
CryptoGPT recently rolled out AI assistant "Alex" and is developing its ZK rollup layer 2 blockchain and a data-to-AI engine, which collects, encrypts and transfers data for commercial applications. CryptoGPT's broad aim is to allow users to earn money by monetizing their data across fitness, dating, gaming and education.
What caught our eye? 👀
Social Future's product, MAY ("Meet Another You"), enables immersive social experiences via personalized avatars. Built with blockchain and AI, MAY allows users to generate ownable 3D virtual worlds where avatars convey real-time expressions and voices. I guess neither web3, nor the metaverse is dead.
What caught our eye? 👀
Teton.ai is a startup in the health tech industry, originating from Denmark, which has designed an AI-based solution to ease the burden on nurses. Their AI assistant offers a comprehensive overview of the ward’s events and alerts nurses when patient care is required, monitors sleep, and issues fall warnings. Nurse!!! Sorry… NurseGPT!!!
New developments to spam your #random Slack channel 💬
🏆 A third of the AI companies featured in this year's Forbes AI 50 are using generative AI
🩻 AI company Optic has launched AI or Not, a web service that attempts to identify whether an image was generated by AI
💰 OpenAI announced that it would offer rewards to people for identifying bugs in ChatGPT and its other AI services
💸 Stability AI is struggling with finances
🖼️ OpenAI looks beyond diffusion with ‘consistency’-based image generator
🔒 The US government is looking into options of AI regulations
🐣 Elon Musk is moving forward with a new generative-AI project at Twitter after purchasing thousands of GPUs
Things to learn when you need a raise
How to use chatGPT in your PM work?
Building a ChatGPT Plugin for Lex Fridman Podcasts
A Recipe for Training Good Generative Models
Want access to the Generative AI investors database? Share and subscribe to AI In the Middle (below) to receive it! ⬇️
Before you go
The future of video looks pretty slick.
The advancements happening with VIDEO in the Generative AI space is going to be one of the most exciting things to follow in 2023.
Progress has entered an exponential growth curve.
Here's what to keep an eye on:
— Roberto Nickson (@rpnickson)
7:37 PM • Apr 6, 2023
And scene. That’s all for today! Thank you for always reading, sharing, and subscribing. Want to share something with us? Slide in our DMs, they’re open.
— Calin Drimbau (@calindrimbau)