Oct 18, 2020
More accurate completions
We've been putting a lot of time into our prompt design, which directly increases the quality and efficiency of the completions we offer. This is paying off well and we are looking forward to sharing some results with you during the upcoming weeks.
Way faster testing environment
From the start, we decided to develop Flowrite in steps, increasing the challenge level gradually. First, making the engine functional in an isolated environment and then, raising the bar by taking it into the real world – use-case by use-case.
During the past two weeks, we have made some core improvements to the testing environment, including the capability to dynamically control the brains that produce completions. We also took the first steps towards making Flowrite learn to adapt by itself.
Last week, we headed to the Finnish countryside to launch Flowrite after working on it for 5 weeks. The launch was successful with hundreds of people on the waitlist during the first 24 hours. Read the whole story here.
- Following our launch, there has been quite a bit of interest to 1) see the product in action, and 2) learn how we got access to GPT-3. So, here are answers to both:
- Want to hop on board to build the future of writing? We opened a signup form on our careers-page to get you notified when we start hiring.
Share the update