Humanoid Robots Are Now a Reality

Over the years, we have seen humanoid robots such as ASIMO, Pepper, Sophia, and others talk to humans and perform certain tasks (mostly for entertainment). The latest Figure 01 (powered by OpenAI) seems to have more advanced capabilities than ever before—even leaving Tesla’s Optimus and Boston Dynamics’ Atlas speechless.?

And that too in a short span of less than two years.?

This week, the video of ‘OpenAI’s ChatGPT Getting a Humanoid Body’ shocked the world. “As you can see from the video, there’s been a dramatic speed-up of the robot, we are starting to approach human speed,” said Figure robot founder Brett Adcock, saying that the video shows end-to-end neural networks.?

But, how did it achieve this? Adcock said that Figure’s onboard cameras feed into a large vision-language model (VLM) trained by OpenAI. “Figure’s neural nets also take images in at 10hz through cameras on the robot. The neural net is then outputting 24 degrees of freedom actions at 200hz,” he added.??

This explains how Figure 01 could seamlessly differentiate between thrash and apple. Leveraging OpenAI’s advanced vision-language model (most likely GPT-5 with Vision), the robot showcased improved reasoning and conversational abilities. “I think I did pretty well. The apple found its new owner, the trash is gone, and the tableware is right where it belongs,” said Figure 01, showcasing the reasoning capabilities of its actions.?

Just two months ago, Figure 01 made coffee only using neural networks. “This is a full learned, end-to-end visuomotor policy mapping onboard images to low-level actions at 200hz,” highlighted the team.?

From these accomplishments, Figure is surely emerging as the next hottest AI and robotics startup, following in the footsteps of OpenAI.

“2024 will be the year of Embodied AI,” said Adcock, believing that advanced AI capable of complex tasks will likely develop in parallel with, or even slightly ahead of, reliable humanoid robot hardware.

Explore how Figure is accelerating the evolution of humanoid robots here.?

Join AI Forum for India: Our Discord Community for AI Ecosystem, in collaboration with NVIDIA. Here>>

Missing NVIDIA GTC 2024 Would be a Foolish Sin

One of the most awaited, world’s premier technology conferences, NVIDIA GTC 2024, is happening next week (March 18-21, 2024), at the San Jose McEnery Convention Center, in San Jose, CA, you surely don’t want to miss.?

Click here to know more.?

Top Stories of the Week >>?

GenAI is Going to Change India’s Agriculture Forever

India’s agricultural sector, which employs about 65% of India’s population, has been grappling with a significant knowledge gap, hampering the transfer of modern farming practices and technologies to millions of farmers across the country, according to experts and recent reports.

To address this challenge, the next generation of generative AI startups, including KissanAI, Wadhwani AI and Gooey.AI, alongside initiatives such as Jugalbandi and Ama KrushAI, are helping farmers make informed decisions.?

Read more here.?

Devin is Just Canva for Developers

Devin, the world’s first AI software engineer, is democratising coding like never before, bringing ease and efficiency to developers and non-developers, similar to how Canva simplified design for designers and non-designers.?

Learn more about Devin here.?

People & Tech >>?

UiPath is Built With Indian Talent, Blood and Effort

India is emerging as one of the fastest-growing markets for UiPath, a leading global software company specialising in robotic process automation (RPA). Notably, UiPath's initial customer was Sutherland, a business process transformation company based in India. Discover more about UiPath’s connections to India and its early beginnings here.?

Meet the Creators of ????????

A team of Indian researchers based in the US—comprising Aakash Patil, a postdoctoral researcher at Stanford University; Mrunmayee Shende, cofounder of CourtEasy AI; and Niraj Kumar Singh, a machine learning engineer at Inbound Health—has recently developed MahaMarathi 7B. This new model joins the league of Indic LLMs, including Kannada, Telugu, Malayalam, Tamil, and Odia Llama.?

Learn more about MahaMarathi here.

Hackathons >>?

Bhasha Techathon, where Innovation Converges with Impact!

MachineHack, in collaboration with Digitial India Bhashini Division, and Google Cloud, has launched Bhasha Techathon, a first-of-its-kind hackathon to build effective and indigenous solutions to language-specific challenges.?

Register now.

AIM Videos >>?

[Must Watch] AI Jobs for Indian Villagers and Small-Town Graduates

In our latest episode of Story Kya Hai, AIM attempts to remove the stigma around job security due to rapid AI advancements and touches upon some of the latest innovations that are creating new employment opportunities, particularly in smaller towns and cities.?

AIM Shots >>?

• Agri-tech startup KissanAI launched Dhenu, a series of fine-tuned vision language models (VLMs) to revolutionise disease detection in crops.?
• MGX, the newly launched AI-focused investment company in Abu Dhabi, is in talks about investing in OpenAI’s chip venture.?
• Staqu Technologies launched Crime GPT, an advanced AI tool that allows police to access information on criminals rapidly from a massive digitised database.?
• AI4Bharat recently created IndicLLMSuite, a collection of resources for Indic LLMs covering 22 languages with 251 billion tokens and 74.8 million instruction-response pairs. Learn more here.?
• Apple recently acquired a Canadian AI startup, DarwinAI, which specialises in visual quality inspection offering manufacturers an end-to-end solution to improve product quality and production efficiency.?
• Oracle recently announced the integration of new generative AI capabilities into its Oracle Fusion Cloud Applications Suite.
• Cerebras Systems and G42 announced the development of Condor Galaxy 3 (CG-3), the latest addition to their AI supercomputing constellation. Read more here.?
• Andrew Ng’s DeepLearning.AI recently launched a free short course called Knowledge Graphs for RAG in collaboration with Neo4j.?
• Google researchers recently introduced Synth2, a novel approach for training Visual-Language Models (VLMs) using synthetic image-text pairs.
• Google DeepMind recently introduced the Scalable Instructable Multiworld Agent (SIMA), a generalist AI agent for 3D virtual settings.?
• Google Sheets users can now use Claude 3, Anthropic’s latest LLM.?
• Microsoft recently announced the general availability of Microsoft Copilot for Security, set to debut on April 1, 2024. Check it out here.?
• OpenAI’s CTO, Mira Murati, recently announced that the company will make its latest text-to-video generator, Sora, publicly accessible later this year.?
• Yann LeCun, the chief of Meta AI, recently appeared on a podcast with Lex Fridman and said he recently met with an Infosys founder who is funding a project based on Llama 2.?