146 | 👾 🧰 Meet Q* (or why OpenAI CEO was fired?)

In today’s Brainyacts:

1. Project Q* primer

2. Talking with YouTube videos

3. Lawrence, the AI paralegal, and more AI model news

4. The “Napster” lawsuit of the AI era and other related content

👋 to new subscribers!

To read previous editions, click here.

Lead Memo

👾 🧰 “Is this a tool we’ve built or a creature we have built?”

This is a direct quote from Sam Altman, OpenAI recently reinstated CEO. He said this a day before he was fired from the company last week. Was he being rhetorical or literal? Hmm . . .

Meet Project Q*(pronounced “Q star”)

Background

OpenAI recently faced internal turmoil, leading to the temporary dismissal of CEO Sam Altman. The root cause of this upheaval seems to be linked to a new AI model known as Q*, developed under the leadership of OpenAI's chief scientist, Ilya Sutskever.

What is Project Q*?

Project Q* represents a significant advancement in AI capabilities. Unlike existing models, Q* reportedly has the ability to solve basic math problems, a task that has been challenging for current AI technologies. This ability signifies a potential leap in AI's capacity to reason logically and abstractly.

Combination of Techniques

Sophia Kalanovska, a cofounder of AI-training startup Tromero, suggests that the name Q* implies a fusion of two AI techniques: Q-learning and A* search. This combination could meld deep-learning methods, like those powering ChatGPT, with rules programmed by humans. Such an approach might address the issue of AI 'hallucinations', where AI produces nonsensical or unfounded outputs.

Towards Artificial General Intelligence (AGI)

Experts believe that Q*'s advancements could be a stride towards AGI, or superintelligence. AGI refers to an AI's ability to understand, learn, and apply its intelligence to a wide range of problems, akin to human cognitive capabilities. Q* seems to bridge experiential learning with factual reasoning, a key aspect of human-like intelligence.

It attempts to solve a crucial problem in AI: the ability to generalize and solve unseen problems. This capability, particularly in mathematics, is a significant marker of intelligence. Current AI models, including those like GPT-3 or GPT-4, have limitations in generating new ideas or solving novel problems without relying solely on their training data.

Controversy and Concerns

The development of Q* has sparked concerns within OpenAI. A group of staff members wrote to the board, warning that Q*'s breakthroughs could pose a threat to humanity. This alarm, along with other factors, contributed to the temporary dismissal of Sam Altman. The controversy revolves around the potential risks associated with rapid advancements in AI, especially without fully understanding their implications.

Implications and Future Directions

The development of Q* has highlighted the dual nature of OpenAI's mission: balancing the pursuit of groundbreaking AI advancements with the ethical and safety concerns of such technologies. The incident underscores the need for careful consideration and possibly regulation of AI development, especially as it moves closer to AGI.

Conclusion

Project Q* stands at the forefront of AI research, symbolizing both the immense potential and the profound challenges of creating AI that mirrors human cognitive abilities. While its full capabilities and implications are yet to be fully understood, Q* marks a significant milestone in the quest for AGI, sparking both excitement and caution in the AI community and beyond.

Whether or not the supposed hype around Q* is real or not, we don’t know yet as OpenAI hasn’t shared much more on it. There are skeptics and there are believers. Just another day in the age of AI.

Spotlight

Use Google Bard to summarize YouTube videos

YouTube generative AI will soon let you have a conversation about a video you’ve just watched. Google is testing a new conversational AI tool for YouTube that can answer questions about videos, provide recommendations, and more.

Below is a snapshot of what I did. I simply asked a question about a long YouTube video I had watched. I want the key points from it.

In the near future, YouTube videos will have an ASK button that you can use to begin interacting with the video content - beyond asking questions about the content, you can begin a conversation about it.

This tool lets you get answers to questions about the video you’re watching, recommendations for related content, and more, all without interrupting playback. For certain academic videos, the tool can aid learning journeys by providing quizzes and responses that encourage deeper understanding. If you’re a part of the experiment, you can access the tool by tapping ✨Ask beneath select videos and begin by asking questions about the video or choosing a suggested prompt.

AI Model Notables

► OpenAI ChatGPT Voice Conversation is Now Available to Everyone Free

► OpenAI brings Sam Altman back as CEO less than a week after he was fired by board

► ChatGPT alternatives fast catching up: Latest Grok, Claude, Bard, Ernie updates prove so

►Yes, I just upgraded my Twitter/X subscription to Premium +. Why? Grok will be available to all + subscribers next week

► Lawrence the ‘AI paralegal’ passes Solicitors Qualifying Exam (SQE) with flying colours

► AI startup Inflection AI just announced Inflection-2, a HUGE new 175 billion parameter language model — and claims its capabilities exceed Google and Meta's top models and “is very close” to catching GPT-4. Inflection-2 will soon be powering Pi

News You Can Use:

➭ The lawsuit that could rewrite the rules of AI copyright. "We’re in the Napster-era of generative AI, says Butterick, with piracy fueling innovation"

➭ US 5th Circuit Court seeks regulation on lawyers’ AI use in legal filings

➭ Slaughters seeks out generative AI-focused products for next incubator program: Magic Circle firm opens applications for fourth round of its Collaborate initiative. The 12-week program is open to all types of legaltech products at all stages of development

➭ Risk Outlook report: The use of artificial intelligence in the legal market

➭ United States: FTC Signals Additional AI-Related Enforcement Is Coming

➭ New law school study says students who use generative AI finish work quicker

➭ Meet the first Spanish AI model earning up to €10,000 per month

Who is the author, Josh Kubicki?

Some of you know me. Others do not. Here is a short intro. I am a lawyer, entrepreneur, and teacher. I have transformed legal practices and built multi-million dollar businesses. Not a theorist, I am an applied researcher and former Chief Strategy Officer, recognized by Fast Company and Bloomberg Law for my unique work. Through this newsletter, I offer you pragmatic insights into leveraging AI to inform and improve your daily life in legal services.