On a Scale From One to Ten, With No Sevens

Scoring a 7 out of 10 in, well, pretty much anything is quite good, you know? Not super amazing, but definitely better than just okay. What a nice, comfortable spot to be in. Fuck 7s.

We’ll also explore nine exciting new startups to understand where some markets might be going in the next few years, thanks to the current boom in AI technology.

Quick hits

• Slingshot is working on the next wave of video games, where characters and gameplay are driven by generative AI. This concept sounds really cool! I expect that future waves of indie games will use mechanics like these, much like how they currently do with randomly generated maps.
• Smobi is a marketplace designed to simplify the process of purchasing businesses valued at $5 million or less by overseeing the entire transaction directly through their platform. If Opendoor can revolutionize real estate sales and OpenStore can manage Shopify shops, then why not provide a similar service for small and medium-sized businesses?
• Fragment integrates human supervision into automated processes, an approach known as human-in-the-loop. While the current wave of generative AI is getting smarter by the day, it’s getting clear that it isn’t capable of replacing most human jobs. However, what AI excels at is quickly and automatically evaluating large data sets. Fragment enables human agents to step in to handle fallback or review tasks when the AI encounters errors.
• Flint lets teachers leverage AI for personalized learning experiences by creating chatbots to tutor or assess students through one-on-one conversations. With Flint, teachers can establish rules for the AI, specify reading levels, and set grading criteria. The AI then adapts to each student’s level and provides assistance when they encounter difficulties. When UX research shows that Bing Chat's responses are calibrated at a 13th-grade reading level, similar to what a university freshman might encounter, while ChatGPT responds at an impressive 16th-grade reading level, there’s a clear need for a tool that kids can use.
• Casehopper is streamlining the immigration process using AI. It’s clear that the legal industry is ripe for some disruption due to LLMs, even though changing how lawyers operate will undoubtedly be challenging—as it always is. It’s reasonable to expect that certain legal workflows, such as visa processing, will become more efficient with the latest technology.
• Patents can be another similar example. Solve Intelligence uses AI to help attorney write patents.
• Corgea helps businesses fix vulnerable code, reducing engineering costs and time by 80%. Through AI-driven code fix generation, it streamlines the process, allowing engineers to approve fixes efficiently. Security is often not a top priority for startups until they grow bigger. But with AI helping out, that might change.
• And Contour does the same for QA and end-to-end UI tests. The main trend in the coming years might be using big models like this one for specific needs. People who know a lot about certain problems will tweak these models to work better in different industries and situations.
• Sideko uses OpenAPI to make SDKs for your API that really work. This is a great idea because making libraries can be a hassle for developers. If it’s automated, it’s a big win for engineering teams eveywhere.

🙏 My work is reader–supported. You can get a membership here!

📣 Before Growth has grown through word of mouth. Want to help? Share it on Twitter here, Facebook here, or LinkedIn here.

📚 My new ebook Generative AI in Product Design offers case studies on AI and just enough theory for you to build your next app with gen AI. Get your copy here!

Today’s topic

Imagine this: Your HR department wants to run a pulse survey to gauge your team’s feedback on a regular basis. Now it’s your turn to contribute. They ask a question like, “How likely are you to recommend your company as a place to work to people you know?” Your experience at work has been generally positive—not great, not terrible—but there have been a few bumps along the way. Though maybe it’s just you…? In the end, you opt for a respectable 7 on the ten-point scale.

On another day, your manager asks you to rate them in a performance review. Despite having reservations about how they handle feedback when under pressure but, after all, who doesn’t get defensive when cornered? Wanting to avoid unnecessary drama, you settle on another solid 7.

It's a few days after the new year. As you try to pick up the habit of journaling with a shiny new app, it gives you a writing prompt that asks you to rate your day and explain your reasons. Deep down, you sense that things have been a bit off for a while now… but perhaps it’s just the winter blues? The cold and darkness could easily be to blame. Besides, you pride yourself on being a positive person with a strong sense of agency. You convince yourself that things haven’t been too bad after all. In the end, your rating lands on a 7 once again—there’s still hope for improvement, right?

Do you see where I’m going with this?

A seven often is as a non-answer, crafted to avoid causing offense to anyone—including yourself. This is perfectly normal. We humans are social creatures. We want to maintain harmony and avoid upsetting others without a reason. Even when there is cause for criticism, we may still go to great lengths to sidestep confrontation if we don’t deem it absolutely necessary. This approach is not a bad day-to-day strategy; it often fosters cohesion in groups.

However, there are times when prioritizing truth-seeking is more important than maintaining cohesion. In each of the three scenarios outlined earlier, it would better if individuals were compelled to choose the confrontational option. The workplace cannot progress if feedback on issues is withheld. A manager cannot address and improve their behavior if they remain unaware of its impact. And if you convince yourself you’re fine when you’re not, you may delay the acknowledgment of the need for change.

So next time you design a system meant to uncover issues, have people rate stuff on a scale from one to ten—but with no sevens.

Eliminating the “safe” option will prompt them to consider which extreme they lean toward—are things good or bad? This encourages everyone to recognize that it’s fine to feel dissatisfied and that raising concerns is expected. Hopefully, it also creates an environment where problems are addressed in a peaceful and productive manner when brought to light. However, that’s a separate matter altogether, and it’s important for you not to panic when people finally start telling you the truth.