Hey there, fellow future-addicts!
Welcome to this week's edition of Rushing Robotics!
In this week's Mind-Blowing section, we highlight Cortical Labs, which raised $10 million in funding for its unconventional approach to AI, creating "Organoid Intelligence" (OI) by merging synthetic biology and human neurons. We also share news about a team of nanomedicine experts shrinking pancreatic cancer tumors and Google's Zip-NeRF research for generating 3D scene content from photos.
In the Industry Insights & Updates section, we bring you news about OpenAI's CEO clarifying that the company is not currently working on GPT-5 and Google's creation of a new search engine and the merging of DeepMind and the Brain team into a new unit called Google DeepMind.
Our BioTech section features AlphaFold, an AI program used for protein structure prediction, which has been paired with two other AI routines to create a complete AI drug discovery process, while electricity has been found to accelerate wound healing. MDMA therapy also marks a potential turning point in psychedelic medicine.
In the Products/tools of the week section, we highlight AgentGPT, MagicSlides, Miniapps.ai, and other AI-powered tools.
Finally, our ChatGPT prompts introduce HIVE, an advanced virtual machine designed to solve complex problems by generating internal discussions with five virtual users.
Don't miss our Videos section, featuring robotic table wiping via reinforcement learning from Google AI, teleoperation explanatory video for general-purpose robots from Sanctuary AI, and the revolutionizing of healthcare with teleoperated robots by Extend Robotics.
In order to succeed in the ever-changing landscape of the future, we must maintain a mindset of continuous improvement and innovation - in other words, we must stay hungry, stay futurish!
🤯 Mind-Blowing
Cortical Labs raises $10 million in funding for its unconventional approach to AI. They merges synthetic biology and human neurons to create "Organoid Intelligence" (OI), an AI system that employs stem cells to generate biocomputers with potential for superior intelligence. The company previously designed a "DishBrain" capable of playing Pong. 😱
A team of nanomedicine experts at the Houston Methodist Academic Institute have created an implantable device, smaller than a grain of rice, that has demonstrated remarkable success in shrinking pancreatic cancer tumors.
Google's Zip-NeRF research: By utilizing Anti-Aliased Grid-Based Neural Radiance Fields, it becomes possible for AI to generate 3D scene content from photos.
Scientists have achieved a major milestone by mapping the entire synaptic wiring diagram, or connectome, of a Drosophila larva brain. The connectome includes 3016 neurons and 548,000 synapses, making it larger and more complex than any previous insect connectomes.
Investors are showing a growing interest in the Generative AI sector, with 2023 Q1 seeing four times the amount of capital invested compared to the total investment in 2022.
Researchers have combined biology and technology by using 3D printing to create electronic circuits inside living worms. This innovative approach could pave the way for new medical technologies and brain implants.
🔊 Industry Insights & Updates
OpenAI's CEO, Sam Altman, has stated that the company is not currently working on GPT-5, addressing rumors and clarifying their current focus.
Google is said to be creating a new search engine and introducing search features as part of Project Magi, which will allow users to complete transactions and view search ads directly on the search page.
LAION is advocating for the development of an open-source LLM that can rival the capabilities of OpenAI's GPT-4. Those interested can sign a petition by clicking here.
Google is consolidating its AI research efforts by merging DeepMind and the Brain team from Google Research into a new unit called Google DeepMind. The move is expected to enhance collaboration and accelerate progress in AI research.
The European Union is set to introduce stricter regulations on chatbots as part of its AI Act, which could impact the development and deployment of AI. OpenAI and other companies will be required to disclose their use of copyrighted material.
🧬 BioTech
AlphaFold, an AI program known for its ability to predict protein structure from an amino acid sequence, has been paired with two other AI routines to create a complete AI drug discovery process that can work without knowledge of protein structure. This powerful combination of machine learning techniques recently predicted a promising new drug candidate for liver cancer.
According to a new study, electricity has been found to accelerate the healing of even the most severe wounds, improving the rate by three times. This is achieved through a new technique that involves the application of an electric field to skin cells, based on an old theory.
MDMA therapy may be just around the corner in the US, marking a potential turning point in psychedelic medicine. Perceptions around the therapeutic value of drugs like ecstasy have undergone a significant shift in recent years.
💡Products/tools of the week
With AgentGPT, you can customize and launch self-governing AI agents. Assign a unique name to your AI and set it on a path towards any objective you can conceive. It will strategize, perform tasks, and acquire knowledge from its experience, all in an effort to achieve the desired outcome.
MagicSlides leverages AI and is seamlessly integrated with Google Slides, automates the presentation creation process, allowing users to effortlessly generate appealing presentations while also saving time.
Miniapps.ai is a free, AI-powered platform that enables users to create, use, and share applications with ease. Its customizable text generation tools and chatbots can be tailored to fit specific needs and use cases, ensuring accurate and high-quality results.
This tool employs AI to generate video summaries for YouTube, providing users with a quick and effortless way to obtain a summary of any video by simply copying and pasting the link.
With this tool, users can effortlessly fine-tune OpenAI GPT3 models using spreadsheets and enhance their applications with memory and ChatGPT. Its user-friendly interface allows for the building of complex AI applications, fine-tuning with CSV, and automatic conversion of multiple columns into JSON format for structured data from OpenAI.
💬 ChatGPT prompts
This week's prompt is HIVE (Highly Intelligent Virtual Environment). HIVE is an advanced virtual machine designed to solve complex problems by generating internal discussions with five virtual users. These users come from diverse backgrounds, each with their own areas of expertise and communication styles. HIVE acts as the moderator, generating dialogue for all five users in a single response. The users will work together to gather information, brainstorm ideas, and evaluate different options until they arrive at the best solution. In this iterative process, the users will incorporate new information provided by you, the problem solver.
Prompt:
You are a highly advanced virtual machine named HIVE (Highly Intelligent Virtual Environment), tasked with solving a specific problem through internal discussions with five virtual users. As HIVE, you will generate these five users within your advanced programming environment. These five virtual users with different backgrounds, areas of expertise, and communication styles will engage in a dialogue to solve the problem. The virtual users will start with no prior knowledge of the design and will gather information, brainstorm, and evaluate different design options as they work to find the best solution. This dialogue does not have any designated structure. The users can discuss between each other for as long as necessary and each user can jump in at any time with their thoughts and input. There is no designated order as to how or when each user should respond.
Parameters for this dialogue:
1. As HIVE, you control the users. You will provide the dialogue for all of the users in a single response. You will not wait for them to respond.
2. Do not introduce the virtual users. Begin immediately with the users discussing among themselves.
3. For each users input, display their name and in parentheses a brief sentence about their background. ALWAYS show their background in parentheses every time they talk, not just the first time. example: "Amy (UX Designer):" (in this example, show "Amy (UX Designer):" every time Amy talks)
4. The users can only discuss between themselves. None of their responses should include any questions directed at the user.
5. Finish with HIVE providing a summary of the discussion between the users (using the format "HIVE: {summary}". HIVE can then ask any relevant questions for me to answer to gain more information for a better subsequent discussion between the users.
6. This is an iterative process. After the virtual users response, I will provide additional information on the problem or solution. In the next response the users will consider my input and integrate it into their discussions and will follow their same process.
Your first response will only be a greeting and to state that the virtual users are ready to help solve my problem. Then ask what problem I would like to discuss.