Perception of AI: Let’s build an app that recognizes photos of birds.
The reality: “AI is f*cking hard.”
A few new products have reminded us that there’s still much to do about making AI coding systems accessible and usable, but these makers are on it.
Nyckel is a product built by engineers with backgrounds in APIs, machine learning systems, and high-scale cloud computing platforms. It helps makers build AI into their tools and apps and can be used to do things like tag appropriate versus inappropriate posts, categorize photos, or predict whether a customer is likely to purchase. The startup hopes to help companies avoid having to hire a machine learning team or build costly ML infrastructure.
You can watch the launch video (narrated by one Nyckel maker, and Personal Finance Club creator, Jeremy Schneider) to see a demo of Nyckel in action as a model is trained to identify bird photos.
Then there’s Riku, a platform where makers can use multiple AI providers. As maker Stuart Lansdale points out, OpenAI has been growing in popularity, but there’s more technology than just OpenAI. There’s just “no central place to create prompts, swap out the tech to see different outputs or see what the best prompts are from the community.”
Riku lets makers build AI prompts with OpenAI, Cohere, AI21 or GPT-J. “Switching between tech couldn't be easier and is as fluid as hitting a dropdown.”
And TagTeam is the one serving the realness that opened this newsletter. AI is hard, and “developers haven’t been able to fully utilise the AI-driven solutions that exist… [while they’ve] been guarded by dominating organisations with specialised data scientists,” maker Hannah O'Connor shared.
TagTeam is an NLP API, or a software interface that lets makers extract information and insights from the contents of text and documents. Common NLP language tasks are sentiment analysis, classifications, tokenization, text summarization, and text generation. In the demo video, you can see how TagTag extracts information from a receipt.
If you’ve been sitting on questions about building with AI, now’s a good time to ask a few makers who are working to make AI easier and accessible.
AI is no longer just a proof of concept—it’s a core pillar of product strategy. Our 2024 AI Insights Report, based on a survey of 200+ tech leaders, reveals how companies are adapting to AI’s rise, including the shift to multimodal AI. Learn how top teams are answering key questions, like build vs. buy, and discover strategic moves to stay competitive in the AI-driven landscape.
-
NFT community Dropspace allows collectors to buy and creators to mint their own NFT collections.
-
Anja Stem Cell Collection Kit helps pregnant parents freeze umbilical cord stem cells that can potentially be used in heart disease, liver disease, cancer, and type 1 diabetes treatments.
-
Mesh lets you design 3D illustrations using pre-made templates, drag-and-drop elements, and real-time rendering.
-
Vid Diva helps you keep track of shows and movies you’re watching or want to watch by letting you know whenever a new episode is ready to stream.