43% of Workers Lack AI Guidance

If you’re new here, this is the latest edition of the Build Mode newsletter, where we gather the collective wisdom of the people building with AI, designing the future of work, and leading the most important companies of the next decade. Subscribe here to get the top insights in your inbox every week.

Big Idea: Should we just stop talking about prompt engineering?

At the bleeding edge of AI research, there are some really interesting things happening with prompts. A recent paper by one of our favorite AI experts, Ethan Mollick, found that using a political thriller-themed prompt can get better answers on a math problem set.

”You have been hired by important higher-ups to solve this math problem… The life of a president's advisor hangs in the balance.“

Mollick calls this the “irreducible weirdness” of prompting: how unexpected tweaks to inputs can improve outputs from large language models.

But there’s a flip side to the over-emphasis on prompting in the discourse. It can make generative AI seem scary and inaccessible for new users.

As of January 2024, one in four desk workers have tried generative AI, according to Slack’s latest report. That’s up from one in five back in September. But it’s still pretty low.

Connor Grennan, the Dean of Students at NYU Stern, hates the term prompt engineering. He believes it feeds into the sense that you need to have specialized knowledge to use ChatGPT effectively.

“I personally talk to so many people struggling to figure out how to get started,” Grennan wrote on LinkedIn. “They think they need to learn something special. Which means they are not taking advantage of this technology that can genuinely improve their lives.”

But there’s no skill necessary to get started with ChatGPT, or even to master it. The learning curve is basically flat. You just need to be clear on two things: 1) what you already know and 2) what you are trying to achieve.

Then you just talk to ChatGPT like a real person. Don’t get crazy with the requests. Keep it simple. Iterate on your answer. And, unless you’re going pro, stop worrying about prompt engineering.

If you’re struggling to get your team to adopt generative AI tools, this is the message that we would use. You know how to ask for what you want, so you already know how to use AI tools. Now, get in there and have some fun.

Missed last week’s issue of Build Mode? Read it here.

CHART OF THE WEEK

Women are closing the gender gap on AI adoption

AI bros look out. The gender gap in AI usage is closing fast, with women accelerating their use of AI tools faster than men, according to Slack’s latest survey.

The survey also found that 43% of workers haven’t received guidance on how to use AI tools at work. Of those who have used AI tools, 80% say this technology is already improving their productivity.

“The majority of people who are using AI and automation are already starting to experience productivity gains,” said Christina Janzer, Senior VP of Research and Analytics at Slack. “But the data indicates that failing to provide guidance or instruction on AI may be inhibiting your employees from giving it a try.”

This reinforces a theme we’ve been exploring here at Build Mode, which is that AI adoption is ultimately a change management problem. For Women’s History Month this March, we’ll highlight some of the women who are tackling that exact challenge.

WOMEN IN AI

3 tips for young women getting started in tech

This month we’re spotlighting the badass women building in tech and AI. First up: Jackie Berardo. Berardo is a researcher at Meta and a writer for Forbes. She came by the A.Team Clubhouse in Union Square and gave us three tips for young women trying to break into a male-dominated space:

1. My number one piece of advice is to get a basic familiarity with programming languages like Python. This allows you to better communicate and relate with engineering teams, bridging potential divides. The programming language R is really useful for statistics and data analysis has done wonders for me just being able to speak with the engineering team.
2. Proactively position your diverse perspective and background as an asset when pitching ideas or research. The increased diversity can lead to better, more inclusive products. Showcase the fact that, hey, you actually have access to this population that’s growing and becoming more and more relevant on the internet and in technology. After all, women drive 85% of consumer spend.
3. Leverage and grow community support networks with other women in the field. In general spaces that are more lucrative tend to have less woman. AI is a very, very highly concentrated male space. I think that for a lot of ambitious women in my generation, we've really had to band together and use each other as a resource.

AROUND THE WATERCOOLER

Altman vs Musk: “Would you open source The Manhattan Project?”

Elon Musk kicked off a philosophical debate in the tech world last week when he sued OpenAI, a company he helped found but no longer has a stake in, alleging that it has violated its original nonprofit mission. The suit set off a wave of hand-wringing among Silicon Valley insiders over the societal implications of increasingly powerful AI systems.

In one corner, investors like Vinod Khosla argued that unfettered open-sourcing of AI tech poses national security risks, asking: "Would you open source the Manhattan Project?" Who says that tech nerds can’t ask culturally relevant rhetorical questions just in time for Oscar season?

Khosla sees the AI race on the global stage—an Us vs Them scenario with China. Marc Andreessen fired back that Khosla is "lobbying to ban open source."

Business Insider said it's getting cattier than a 'Real Housewives' reunion. The clash highlights the paradox of keeping transformative AI locked up versus making it publicly available. Musk point is that contrary to it’s name, OpenAI isn’t very open.

Oppenheimer’s Cillian Murphy, sadly, was unavailable for comment.

Amidst the grandstanding, legal experts say Musk's claim against OpenAI may be a long shot. But the suit could expose OpenAI's inner workings during the discovery process—including its controversial pivot from nonprofit to for-profit under the helm of the recently-ousted and -reinstated CEO Sam Altman. Musk is angling for a trial. So this drama is just getting started.

“I have mixed feelings about Sam,” Musk said at the DealBook conference. “The ring of power can corrupt, and he has the ring of power.”

We love to see a Lord of the Rings reference, but Musk is clearly trying to create a ring of power of his own.

"We believe the claims in this suit may stem from Elon's regrets about not being involved with the company today," OpenAI chief strategy officer, Jason Kwon wrote in an internal memo on Friday, according to CNBC.

Zing.

EVENTS

The Gen AI Salon: The Women Shaping AI

On March 27th, A.Team is bringing together an all-star panel of enterprise leaders to discuss the gender bias issues in AI and reveal breakthrough frameworks and processes that they’re using to accelerate product development.

We have some great speakers lined up, including Salesforce’s Katherine von Jan, Adrian Tatsch from McKesson, Rajamma Krishnamurthy from Microsoft, and ServiceNow’s Devika Bhalla — join us for an evening dedicated to the women leading the AI revolution.

Reserve Your Spot

DISCOVERY ZONE

Generate a mini comic book from any prompt with Hugging Face’s AI Comic Factory.

MEME OF THE WEEK