I bombed algebra in high school. The new learning mode for ChatGpt is my redemption arc

This week, ahead of Tuesday’s release, we sneak peaked a new learning mode for ChatGpt during Zoom’s Openai Press demo.

Switching the “Ask Anything” chat tool button to “Study and Learn” will convert ChatGpt from a simple Q&A assistant to a personalized tutor. Instead of simply spitting out answers, use Socrates questions, tips and step-by-step guidance to promote active learning. Tell them what you want to study, and it will assess your skill level and even remember your progress throughout the chat.

Watching the demonstration, my eyes widened and my breathing became faster. Research modes may be designed with university students in mind, with instructions from the system created with opinions from teachers, scientists and learning experts. However, decades have passed since my own academic experience, and I quickly learned that I would like to use the learning mode. And I knew exactly what I wanted to use it: high school algebra.

My math failures shaped my life

With the remaining mathematics curriculum of the year, I bombed it completely with algebra. I was deeply embarrassed about how difficult mathematics was for me, even with my tutor. It was a stark contrast to how easily I excel in subjects like English and social studies, and I felt like I didn’t really understand how interested I was in mathematics. There were many questions that I couldn’t answer. What is algebra anyway? Why do you learn that in the first place? Why do they come before geometry and trigonometry? What is the point of all these rules regarding x and y?

Considering what I’m writing about AI, it may seem ironic. This regularly interviews doctoral researchers in mathematics who casually refer to linear algebra in conversation. But I always wanted to get it over in high school math.

Yes, there are real concerns about how AI affects student learning, from fraud and overreliance to privacy and bias. New report this week The rest of the world When rural Colombian students became able to access AI, they found that students began to fail their exams.

But for me, the new learning mode in ChatGpt is already becoming very personal. This is a long redemption arc.

Learning mode at your own pace

I started from places I really couldn’t do before, and the learning mode had a refreshing response right away:

“Treat this like the first day of algebra 1. This time, you can always ask questions and move at a pace.”

I was on it: what teeth Algebra, I asked.

Algebra is part of mathematics, stop manipulating numbers and start working on relationships between numbers.

Instead of asking:
What is 3×5?
We ask:
What number X does this equation make true?

It’s like mysterious mathematics. Something is hidden and our job is to solve it.

“Mysterious Mathematics” – Hmm, I loved it. After giving a drop in the “big idea” of algebra, everything from equation solving to graph lines and word problems offered several options.

I want to:
a) Do you have a question about this?
b) Do you want to select the puzzle type to start with?
c) Continue with the intro and slowly build from here?
Your phone!

He said he wanted to solve the equation. Study mode has a pigeon on the right:

🔍Equation Solving: What is the goal?

Equations are like balance scales. In any case, whatever you do on one side, you have to do it on the other side.

This is your first equation:
3x + 4 = 19

x What do you think you should do first or not to start solving X?
(Please guess. We’ll understand it together.)

AI-driven, personalized learning promise

I sat and smiled. Apparently it’s never too late to resolve X. Even for journalists who always thought she was good at words, but thought the numbers were terrible. I spent several hours going through one and two stages of equations. An equation with variables on both sides. Equations with parentheses; equations that combine terms such as equations. I asked a lot of questions along the way, and sometimes I filled up my answers with “Amazing!” and “I never knew that!”

Like all AI use cases, there are trade-offs. The risk of excessive reliance on AI and reduced critical thinking is real. However, I found myself in his recent book that I was leaning towards the view of my colleague Jeremy Kahn. Learning AI: Survival Guide to the Future of Our Superpoweremphasises the promise of personalized learning with AI. He frames the ideas of AI as one-on-one tutors as one of the most powerful educational opportunities.

My late returns to algebra may not have been the target use case for Openai. But who knows – perhaps I’ll make it a calculation.

So, this is the rest of the AI news.

Sharon Goldman
Sharon.goldman@fortune.com
@sharongoldman

News ai

Microsoft has signed the EU’s AI code of practice, but Meta is declining. Microsoft announcement On Thursday it signed the European Union’s General-Purpose AI Code of Practice, which officially considered one of the first major high-tech companies to do so. The move illustrates Microsoft’s support for the EU’s AI governance framework, despite the demand for simplification of what is described as complex regulations. Other companies have also expressed their willingness to adjust along voluntary codes, including Openai and Mistral, but notable holdouts remain. Meta has refused to participate, and Google has not yet made its position public.

Openai is launching Stargate Norway, Europe’s first AI data center initiative. Openai has been released Stargate Norway As part of a broader Stargate program based on the Openai for country initiative launched in May, we will partner with the government to help build our own AI infrastructure, focusing specifically on data centers. Planned for Narvik and supported by Norwegian partners Nscale and Aker, the facility will offer up to 230mW of AI computing power with plans to scale to 100,000 Nvidia GPUs by 2026. It is part of a global footprint that includes agreements with the UK, Estonia and the EU’s AI GigaFactories initiatives following the Stargate UAE.

AI researchers are approaching the job market like the NBA stars. New York Times Today we have a great story about AI Talent Wars. This is a fierce and advantage for companies like Meta, Openai, Google, Microsoft, and other countries to sign NBA superstars, offering nine-digit compensation packages, and recruiting top young AI researchers as NBA superstars engaged in highly public employment battles. Many of these 20-something “AI free agents” look to informal agents and aides, navigate the enthusiasts, not bound by the salary caps like professional sports, and negotiate top deals. The competition is even tackling the tone of sports spectacles, and streaming outlets like TBPN cover notable industry work changes with league trading deadline talent.

The good fortune we have

Salesforce CEO Marc Benioff Why AI agents aren’t leading to mass unemployment– By Jeremy Kahn

Mark Zuckerberg is pouring billions of dollars into AI “Superintelligence.”– By Sharon Goldman

Meta’s Mark Zuckerberg laid out AI vision that surpassed second quarter expectations and brought stocks to a surge– By Amanda Gerout

Why Booz Allen’s CTO uses Generated AI to create your own deepfake videos– By John Kell

Explanation: Billions of dollars for Silicon Valley’s AI have not actually generated returns yet. Instead, here’s why most companies should embrace “small AI”– By Jason Corso

There is a calendar

September 8th-10th: Fortune Brainstorm Tech in Park City, Utah. Apply to attend here.

October 6th-10th: world ai Weekly Amsterdam

October 21st-22nd: Tedai San Francisco. Apply to attend here.

December 2-7: Newlips, San Diego

December 8th: Fortune Brainstorming ai San Francisco. Apply to attend here.

Turn your eyes to AI numbers

52%

This is the number of developers who are not yet using AI agents, according to a new study from software developers at Stack Overflow, a popular online question and answer platform for computer programmers and developers. This study found that AI agents are not yet mainstream. The majority of developers (52%) don’t use agents, don’t stick to simpler AI tools, and a significant portion (38%) have no plans to adopt them.