OpenAI’s Path to Artificial General Intelligence (AGI)

OpenAI, a leading artificial intelligence research laboratory, has outlined a five-level framework to measure progress towards achieving Artificial General Intelligence (AGI). This framework provides a structured approach to understanding the complexities and potential implications of AI development.

Level 1: Conversational AI – chatbots with conversational language

  • Focus: Developing AI systems capable of engaging in natural and informative conversations.
  • Example: ChatGPT, Google Bard
  • Benefits: Revolutionize customer service, education, and mental health support. Improve accessibility to information and facilitate human-computer interaction.

Level 2: Reasoners – human-level problem solving

  • Focus: Creating AI systems that can solve complex problems, requiring reasoning, planning, and learning.
  • Example: AI systems capable of drafting legal documents, conducting scientific research, or developing complex software.
  • Benefits: Accelerate scientific discovery, increase efficiency in various fields like medicine and engineering.

Level 3: Autonomous Agents – systems that can take actions independently

  • Focus: Building AI systems capable of operating independently in complex environments, making decisions, and taking actions.
  • Example: Self-driving cars, robots capable of performing household tasks, or AI systems managing complex infrastructure.
  • Benefits: Transform transportation, improve quality of life, and enhance efficiency in industries like manufacturing and logistics.

Level 4: Innovators – AI that can aid in invention

  • Focus: Developing AI systems capable of generating new ideas and solutions, demonstrating creativity and adaptability.
  • Example: AI systems designing new drugs, creating innovative products, or composing music.
  • Benefits: Drive economic growth, foster innovation, and potentially lead to breakthroughs in fields like art, science, and technology.

Level 5: Organizational Equivalents – AI that can do the work of an organization

  • Focus: Creating AI systems capable of operating as entire organizations, making strategic decisions, and adapting to changing environments.
  • Example: AI systems managing complex businesses, governments, or non-profit organizations.
  • Benefits: Revolutionize governance, economic systems, and societal structures. However, also raises significant ethical and societal challenges.

According to Bloomberg, OpenAI believes its technology is approaching the second level of five on the path to artificial general intelligence. It’s important to note that this framework is a conceptual roadmap and the exact boundaries between levels may be fluid. Additionally, achieving each level represents a significant technological leap and will likely require substantial advancements in hardware, algorithms, and data.

While the potential benefits of AGI are immense, it’s crucial to address the associated challenges and risks, such as job displacement, bias, and the potential for misuse. OpenAI and other leading AI research organizations are actively working on developing safety protocols and ethical guidelines to ensure that AGI benefits humanity as a whole.

References:

https://www.bloomberg.com/news/articles/2024-07-11/openai-sets-levels-to-track-progress-toward-superintelligent-ai?embedded-checkout=true&sref=HrWXCALa

https://www.forbes.com/sites/jodiecook/2024/07/16/openais-5-levels-of-super-ai-agi-to-outperform-human-capability

Figure Unveiled a Humanoid Robot in Partnership with OpenAI

A yet another milestone in the history of A.I. and Robotics!

Yes, I’m not exaggerating! What you could potentially read in a moment would be a futuristic world where humanoid robots can very well serve humanity in many ways (keeping negatives out of the picture for timebeing).

When I first heard this news, movies such as I, Robot and Enthiran, the Robot were flashing on my mind! Putting my filmy fantasies aside, the Robotics expert company Figure, in partnership with Microsoft and OpenAI, has released the first general purpose humanoid robot – Figure 01 – designed for commercial use.

Here’s the quick video released by the creators –

Figure’s Robotics expertise has been perfectly augmented by OpenAI’s multi-modal support in understanding and generating response of visual inputs such as image, audio, video. The future looks way more promising and becoming reality that these humanoids can be supplied to the manufacturing and commercial areas where there are shortage of resources for scaling the production needs.

In the video, it is seen demonstrating the ability to recognize objects such as apple and take appropriate actions. It is reported that Figure 01 humanoid robot stands at 5 feet 6 inches tall and weighs 132 pounds. It can carry up to 44 pounds and move at a speed of 1.2 meters per second.

Figure is backed by tech giants such as Microsoft, OpenAI Startup Fund, NVIDIA, Jeff Bezos (Bezos Expeditions) and more.

Lot of fascinating innovations happening around us thanks to Gen AI / LLMs, Copilot, Devin, Sora, and now a glimpse into the reality of Humanoid Robotics. Isn’t it a great time to be in?!

Meet Devin, the first AI-based Software Engineer

Gen AI enables writing highly sophisticated code for the given problem statement. Developers can already take advantage of that!

What if a full-fledged tool that can write code, fix bugs, leverages online resources, collaborates with human, and solves gigs on popular freelancing sites such as Upwork?!

Is this a fiction? Well, not anymore.

Meet Devin, the first of its kind, AI-based software engineer, created by Cognition Labs, an applied AI labs company that builds apps focusing on reasoning.

The Tech World is already amazed with the capabilities of Copilot which assists in developing code snippets, however, Devin has a unique capability and is a step-up in terms of its features that it can cater to end-to-end software development.

According to the creators, Devin has the following key capabilities as of writing –

  1. Learn how to use unfamiliar technologies.
  2. Build and deploy apps end to end.
  3. Autonomously find and fix bugs in codebases.
  4. Train and fine tune its own AI models.
  5. Address bugs and feature requests in open source repositories.
  6. Contribute to mature production repositories.
  7. Solve real jobs on Upwork!

Scott Wu, the founder and CEO of Cognition, explained Devin can access common developer tools, including its own shell, code editor and browser, within a sandboxed compute environment to plan and execute complex engineering tasks requiring thousands of decisions. 

Devin resolved 13.86% of issues without human assistance in the tested GitHub repositories as per the publication by creators based on SWE-benchmark that asks agents to resolve challenging problems in the open-source projects such as scikit-learn, Django.

There’s sparkling conversation around the globe that AI could kill basic coding skills written by human and recently NVidia Founder talked about everyone is now a programmer thanks to AI. Of course, I think, human oversight is required to refine and meet user’s requirements.

Thanks to Devin, now the human can focus more on complex or interesting problems that requires our creativity and best use of our time. As of now, access to Devin is only limited to select individuals. Public access is still pending. For more info, visit cognition-labs.com/blog

Meta’s Large Language Model – LLaMa 2 released for enterprises

Meta, the parent company of Facebook, unveiled the latest version of LLaMa 2 for research and commercial purposes. It’s released as open-source unlike OpenAI GPT / Google Bard which is proprietary.

What is LLaMa?

LLaMa (Large Language Model Meta AI) is an open-source language model built by Meta’s GenAI team for research. LLaMa 2 which is newly released for research and commercial uses.

Difference between LLaMa and LLaMa 2

LLaMa 2 model was trained on 40% more data than its predecessor. Al-Dahle (vice president at Meta who is leading the company’s generative AI work) says there were two sources of training data: data that was scraped online, and a data set fine-tuned and tweaked according to feedback from human annotators to behave in a more desirable way. The company says it did not use Meta user data in LLaMA 2, and excluded data from sites it knew had lots of personal information. 

Newly released LLaMa 2 models will not only further accelerate the LLM research work but also enable enterprises to build their own generative AI applications. LLaMa 2 includes 7B, 13B and 70B models, trained on more tokens than LLaMA, as well as the fine-tuned variants for instruction-following and chat. 

According to Meta, its LLaMa 2 “pretrained” models are trained on 2 trillion tokens and have a context window of 4,096 tokens (fragments of words). The context window determines the length of the content the model can process at once. Meta also says that the LLaMa 2 fine-tuned models, developed for chat applications similar to ChatGPT, have been trained on “over 1 million human annotations.”

Databricks highlights the salient features of such open-source LLMs:

  • No vendor lock-in or forced deprecation schedule
  • Ability to  fine-tune with enterprise data, while retaining full access to the trained model
  • Model behavior does not change over time
  • Ability to serve a private model instance inside of trusted infrastructure
  • Tight control over correctness, bias, and performance of generative AI applications

Microsoft says that LLaMa 2 is the latest addition to their growing Azure AI model catalog. The model catalog, currently in public preview, serves as a hub of foundation models and empowers developers and machine learning (ML) professionals to easily discover, evaluate, customize and deploy pre-built large AI models at scale.

OpenAI GPT vs LLaMa

A powerful open-source model like LLaMA 2 poses a considerable threat to OpenAI, says Percy Liang, director of Stanford’s Center for Research on Foundation Models. Liang was part of the team of researchers who developed Alpaca, an open-source competitor to GPT-3, an earlier version of OpenAI’s language model. 

“LLaMA 2 isn’t GPT-4,” says Liang. Compared to closed-source models such as GPT-4 and PaLM-2, Meta itself speaks of “a large gap in performance”. However, ChatGPT’s GPT-3.5 level should be reached by Llama-2 in most cases. And, Liang says, for many use cases, you don’t need GPT-4.

A more customizable and transparent model, such as LLaMA 2, might help companies create products and services faster than a big, sophisticated proprietary model, he says. 

“To have LLaMA 2 become the leading open-source alternative to OpenAI would be a huge win for Meta,” says Steve Weber, a professor at the University of California, Berkeley.   

LLaMA 2 also has the same problems that plague all large language models: a propensity to produce falsehoods and offensive language. The fact that LLaMA 2 is an open-source model will also allow external researchers and developers to probe it for security flaws, which will make it safer than proprietary models, Al-Dahle says. 

With that said, Meta has set to make its presence felt in the open-source AI space as it has announced the release of the commercial version of its AI model LLaMa. The model will be available for fine-tuning on AWS, Azure and Hugging Face’s AI model hosting platform in pretrained form. And it’ll be easier to run, Meta says — optimized for Windows thanks to an expanded partnership with Microsoft as well as smartphones and PCs packing Qualcomm’s Snapdragon system-on-chip. The key advantage of on-device AI is cost reduction (cloud per-query costs) and data security (as data solely remain on-device)

LLaMa can turn out to be a great alternative for pricy proprietary models sold by OpenAI like ChatGPT and Google Bard.

References:

https://ai.meta.com/llama/?utm_pageloadtype=inline_link

https://www.technologyreview.com/2023/07/18/1076479/metas-latest-ai-model-is-free-for-all/

https://blogs.microsoft.com/blog/2023/07/18/microsoft-and-meta-expand-their-ai-partnership-with-llama-2-on-azure-and-windows/

https://www.qualcomm.com/news/releases/2023/07/qualcomm-works-with-meta-to-enable-on-device-ai-applications-usi

https://techcrunch.com/2023/07/18/meta-releases-llama-2-a-more-helpful-set-of-text-generating-models/

https://www.databricks.com/blog/building-your-generative-ai-apps-metas-llama-2-and-databricks