AI Weekly Insights #3 - Robotics and AI
Kharee Smith
Success Manager | Technical Instructor | Software Engineer | Music Producer with a passion for Technology, A.I., Music, and Gaming
Happy Sunday!
Welcome back to another edition of "AI Weekly Insights". This week, we're exploring the fusion of robotics and AI, new safety initiatives, and some cool AI tools you should know about.
New here? Get up to speed with the first and second editions. Also, grab the ChatGPT Ultimate Guide. It's free to subscribers and packed with goodies!
For the week of 10/12/23-10/28/23 (P.S. Click the story’s title for more information ??):
What’s New: Boston Dynamics has integrated ChatGPT into its robot dog, Spot, turning it into a talking tour guide.
How They Did It: This was achieved using ChatGPT and other large language models to train Spot’s responses. They gave the robot a speaker, added text-to-speech capabilities, and made its gripper mimic speech like a puppet.
Why It Matters: I actually said, “This is so cool!” out loud while watching the video below. This is a glimpse into the future where robotics and AI merge, possibly paving the way for androids.
What's New: OpenAI is building a new team called Preparedness to make AI safer as it evolves. The team will assess AI models, and will also try to exploit vulnerabilities in the models.
Key Focus Areas:
Why It Matters: Well, Skynet would be bad but this is awesome to see, as last week we talked about AI being used for phishing. The company behind ChatGPT has a high focus on safety. They have even posted a challenge for people to submit their most creative “misuse” of their AI models.
What's New: Google is investing in Anthropic, the artificial intelligence startup founded by ex-OpenAI executives as confirmed in this CNBC article.
Who is This: Anthropic is the developer of Claude 2, a rival chatbot to OpenAI’s ChatGPT. Claude 2 has the ability to summarize up to about 75,000 words, while ChatGPT, by contrast, can handle about 3,000 words.
Why It Matters: Since generative AI is still new, there’s plenty of room for multiple companies to drive innovation and improve the quality of generative AI products and services. Try Claude here
领英推荐
AI Tools of the Week:
AI History Corner
Generative AI is a relatively new and exciting field of study, but Artificial Intelligence has been around since the 1950s. Here are some significant milestones in AI development starting with the GPT series:
In 2017, Google employees published a research paper titled "Attention Is All You Need." This paper introduced the Transformer architecture (the T in GPT), which mimics how humans pay attention to words or features in different contexts.
This paper directly led to the creation of GPT-1 in 2018 by OpenAI. This model was trained on small datasets (mainly unpublished fiction books) to do very simple tasks like question-answering and semantic similarity. OpenAI's GPT-3 was a major breakthrough, with the ability to translate text, answer questions, perform basic math, use new words in a sentence, and more.
GPT-3.5 (a more fine-tuned version of GPT-3 for chatting purposes) was released, followed by ChatGPT, which made it easy for everyone to interact with OpenAI's model. The most current version of OpenAI's GPT model is GPT-4, which is even more powerful than its predecessors. It will be exciting to see what future iterations will be able to do!
Your curiosity fuels this journey into the AI landscape. If this piques your interest, feel free to share it with friends who’d appreciate a weekly dose of AI insights ??. Drop a comment and let me know what you think. Until next Sunday!
Warm regards,
- Kharee