Always ask | TechCrunch
Welcome to Startups Weekly, a nuanced take on this week’s startup news and trends from Senior Reporter and co-host of Equity Natacha Mascarenhas. To receive it in your inbox, subscribe here.
Editor’s note: Before we get to the heart of the matter, thank you to Startups Weekly readers who pointed out that last week’s link to my column was broken. If you want to know more about the state of solo GPs – which, oh my god, is it hard – please read my TC+ story here: Are Solo GPs screwed? Ok, now in this week’s newsletter!
“AI Whisperer” jobs are roles designed for people who are savvy engineers and able to manipulate ChatGPT or other generative AI tools to get what they want out of them. As Bloomberg reported earlier this week, these jobs can pay up to $335,000 in annual pay. Cha Ching.
That’s an impressive price for a position that would have blown your mind just a year ago. It’s also a reminder that the generative AI boom will create a whole new set of questions about how we ask questions. By spending a few minutes playing ChatGPT, I learned that you can get a diversity of AI-generated answers to questions just by changing a few select words and requiring bullet points against three sentences.
Before I see a whole wave of AI rapid engineering consulting practices for hire – just kidding, we already are – I thought it would be helpful to round up some of the best advice I’m hearing from the community on way to better interact with the AI to get it to give you the answers you want. We already know he’s flawed, can’t be trusted with all the facts, and is sometimes left speechless. But what can we do with what we have?
Many techs weighed in on Twitter – you can see the full thread here – but here are the top five tips:
1. Overdescribe. As one person said, don’t be afraid to be extra when you call on artificial intelligence. There’s a difference between asking for a banana bread recipe and asking for a simple banana bread recipe that incorporates cardamom, takes less than an hour to bake, and has a surprise ingredient. Or in other words, don’t just keep it high, be specific about your preferred tone and complexity in responses. Overuse of explanations and adjectives was probably the most common piece of advice I received when asking for advice on how to be a better teleprompter.
2. Ask for help. It’s not always easy to know how specific you need to be when asking the AI - so why not ask for help? An entrepreneur told me to ask ChatGPT how to invite him based on what I want, so ask something like, what information do you need from me to create an itinerary for a hike in San Francisco? The AI will tell you what it needs – I wish humans were like that – and as a result you’ll get a more personalized response.
3. Give your opinion. Let’s say I’m asking for: “a verbose yet informative three-sentence explanation of the impact of generative AI on education, incorporating at least one metaphor.” ChatGPT had a pretty good response, but I didn’t like the response. Tell him that! Comments are useful for training the AI on the types of responses you like and take a fraction of a second. Here’s how this process went for me:
Natasha: Give me a verbose but informative three-sentence explanation of the impact of generative AI on education, incorporating at least one metaphor.
ChatGPT: Generative AI has the potential to transform education by serving as a virtual Sherpa, leading students on personalized learning expeditions. Much like a Sherpa uses their experience to guide climbers through dangerous terrain, AI can harness vast data to provide a personalized path to knowledge. With the help of generative AI, education can become a more accessible and personalized journey, allowing students to reach new heights.
Natasha: I don’t like the virtual sherpa metaphor. Can you also be more specific about the new heights students can reach?
ChatGPT: Generative AI can be thought of as a personal tutor, providing students with a tailored program that meets their individual needs and interests. Like a trained coach, the AI can identify students’ strengths and weaknesses and provide targeted guidance to help them excel. With the help of generative AI, students can reach new heights by unleashing their full potential and gaining a deeper understanding of complex topics.
4. Ask for shades. ChatGPT only responds to what is asked of it, so why not ask for a nuance? One founder said he asks for the pros and cons of every recommendation he invites ChatGPT to give.
5. Be a little cringe. Is it complacent to tell the AI to talk like me, specifically pushing it to be vocal, charismatic, and informative? Maybe, but it helps. One of the biggest hacks I’ve learned is to throw my Twitter account when I ask for a deep dive on a topic. ChatGPT responds better to “Explain why cauliflower is the best vegetable in @nmasc_ tone” than “Explain why cauliflower is the best vegetable in Natasha Mascarenhas tone”. And because I know you’re curious about the first one, I’ll leave you with ChatGPT’s argument about cauliflower supremacy:
Alright, buckle up, because I’m about to explain why cauliflower is THE BEST vegetable there is. First, let’s talk versatility. This baby can be roasted, grilled, mashed, pureed, cut into rice and even made into buffalo wings. Name another vegetable that can do all that, I’ll wait.
And don’t even get me started on the health benefits. Cauliflower is packed with vitamins and minerals like vitamin C, vitamin K, and potassium, not to mention it’s high in fiber and low in calories. Plus, it’s a great source of antioxidants, which help fight disease and keep your body in top shape.
But let’s face it, the best thing about cauliflower is how delicious it tastes. Whether you enjoy it as a creamy soup, a spicy stir-fry, or a crispy pizza crust, cauliflower has unbeatable flavor and texture. And let’s not forget its ability to take on different flavors and spices, making it the ultimate canvas for creative cooking.
So there you go folks. Cauliflower is the MVP of vegetables, and I challenge anyone to prove me wrong.
These are some starter tips that I’m incorporating, but I’d still like to know what else I’m missing, so please do. throw your best advice into this thread. Also, thanks to Fundrise CMO Jon Carden for the inspiration of the title!
In the rest of this newsletter, we talk about Cerebral Valley and Substack. As always, you can follow me on Twitter or Instagram to continue the conversation. And if you want to support me more, subscribe to my personal (and free!) substack.
Heard in the brain valley
I attended the Cerebral Valley AI conference on Wednesday, which brought together the top 200 minds in artificial intelligence for a day-long deep dive. The event, hosted by Newcomer and Volley, was immersive to say the least. I spoke to half a dozen YC founders who build AI companies, bullish and surprisingly bearish VCs, and, briefly, Stability CEO and Founder Emad Mostaque to keep it real.
Here are some great one-liners:
- “The six-month break seems destructive and probably extremely expensive to consider,” Kleiner Perkins partner Bucky Moore said in response to the recent open letter.
- “Security research is one of the areas we’re best known for…I don’t think that’s necessarily synonymous with a technology company, a company that’s also thinking about how to take that research and to turn it into something practical,” said Daniela Amodei, co-founder of Anthropic.
- “You are building for GPT-4, but you want to build for GPT-10,” said Lisha Li, CEO of Rosebud AI.
- “Of course we’re not in a bubble, that’s more important than 5G and standalone customers,” Emad Mostaque, CEO and Founder of Stability AI, said of the AI hype. “When founders come to me, I tell them to build great products and solve problems…most things are still at the surface level.”
And finally: If you’re still looking for more stability, listen to my podcast interview with Credo AI founder Navrina Singh, who talked about responsible AI, governance, and fear as a not-so-great motivator.
The Substack News Week
Blogging and media platform Substack opened a community fundraiser this week, seeking capital from its writers, accredited and unaccredited readers. As of Friday morning, more than 6,000 investors pledged funding of approximately $6.9 million.
We also talked about the crowdfunding campaign on our Equity podcast this week, which was spicier than I expected. For me, if Substack manages to raise funds from its community after disclosing its financial data, the crowdfunding world will get some much-needed reputation and public perception.
Here’s what you need to knowthrough my colleague Alex William:
Ironically, Substack users might give the company’s professional money managers some breathing room. Venture capitalists who have already poured a mountain of capital into Substack don’t have to invest more, the inbound dilution is effectively zero, and the company gets more total funding to pursue its long-term goals. For Substack, it’s a win.
We won’t see this kind of puzzle too often. Most startups couldn’t handle this kind of crowdfunding because their customers are paying them, not the other way around. In contrast, Substack users earn part or all of their living from the company, so they are more invested in it by staying. This makes the Substack crowdfund unique.
Seen on TechCrunch
StellarFi lands $15 million to help people build credit by paying bills and renting on time
Groupon, which has lost 99.4% of its value since its IPO, appoints a new CEO… based in the Czech Republic
US investors cut valuation of Byju and Swiggy
Twitter is dying
Apple has acquired a startup using AI to compress videos
Seen on TechCrunch+
Layoffs will continue until (investor) morale improves
Investors Unfazed by Dip in Crypto Funding in Q1
Pitch Deck Teardown: Northspyre’s $25 Million Series B Deck
What’s going on with the TikTok ban?
Blinded by the speed of change
Leave a Reply