Some great tricks for reducing hallucinations when performing classification tasks w/a dynamic list of categories: engineering.fractional.ai/taming-llm-res…

AI features feel like magic when they're done well. Totally agree with Michel Tricot that Airbyte's new AI assist feature has that magical feel to it. It's going to save so many engineers so much time. Congrats to Airbyte on the launch of Airbyte 1.0!

Check out this feature that Fractional AI built in partnership with the Airbyte team. It's really exciting to be working on valuable AI use cases that actually make it to production. Come join the team! techcrunch.com/2024/09/24/dat…

Our engineering team is excited about the OpenAI announcements today. Here's one reaction that was just posted in Slack. "The slate of new stuff OpenAI released today is all very cool: - Treading into braintrust/langsmith/openpipe territory. Create datasets of input/output

We're hiring our first Account Executive Fractional AI. This should be a dream role for an ambitious salesperson. - The market is huge, and initially you'll have zero territory limitations. - You'll get a bunch of good leads (last quarter we had 37 quality inbounds and that

This has been one of the more interesting AI projects I’ve seen get to production. Using AI to reliably build API integrations is just barely possible. The universe of API doc websites is very broad and difficult to scrape, the APIs themselves vary a lot, and reliably handling

The question I get the most about building with gen AI: how do you deal with hallucinations? This case study shows how Zapier reduced hallucinations by over 80%.

legacy software companies adding an ai chatbot to their product

Private equity is the sleeping giant of the AI boom. fractional.ai/blog/private-e…

Presented to a group of business leaders today about taking AI projects all the way through to production. Couple of takeaways from their questions: - Companies are getting more sophisticated about which AI projects to take on. Less “demonstrate that we do AI things” and more

This case study shows how you go from unstructured to structured data with LLMs. fractional.ai/case-study/how…

A lesson from a recent AI project for Zapier : find ways to automate the creation of your 'ground truth dataset' Instead of manually building an eval dataset, we tapped into Zapier’s historical codebase — pairing resolved tickets with merge requests to automate ground truth

Lot of people think they should start AI workflow automation projects with open-source models to save money. Don’t do it. Unless you’re buying your own GPUs, the biggest cost for developing with gen AI is typically the developers themselves. Limiting them to open-source models

5 AI myths keeping companies from getting AI into production...and what to do instead. The latest from our CTO and Co-Founder Eddie Siegel fractional.ai/blog/5-ai-myth…

Proper scoping can make all the difference between AI projects that stall at POC and those that cross the finish line -- but getting scoping right is hard. That’s why we wrote a guide to scoping your next enterprise AI project. Access it here: fractional.ai/white-papers/h…

Writing off gen AI because it hallucinates makes about as much sense as writing off aviation because bolts sometimes rust. Many people seem to think that hallucinations and unpredictability render gen AI useless. The truth is that you can build reliable systems with unreliable

Fun little experiment I just did during a scoping exercise. Project included the need to recognize images of celebrities using AI, so I tried pasting an image of Tom Cruise into various models and asking "who is this?" The results are hilariously revealing: GPT-4: “I can’t help

All of the hype points to vertical AI startups taking off. But a closer look at market dynamics tells a different story. travismay.medium.com/the-value-of-d…