Topics
Latest
AI
Amazon
Image Credits:thianchai sitthikongsak / Getty Images
Apps
Biotech & Health
Climate
Image Credits:thianchai sitthikongsak / Getty Images
Cloud Computing
mercantilism
Crypto
Enterprise
EVs
Fintech
Fundraising
contrivance
punt
Government & Policy
computer hardware
Layoffs
Media & Entertainment
Meta
Microsoft
Privacy
Robotics
Security
societal
infinite
inauguration
TikTok
Transportation
Venture
More from TechCrunch
Events
Startup Battlefield
StrictlyVC
Podcasts
Videos
Partner Content
TechCrunch Brand Studio
Crunchboard
Contact Us
Just about everyone is attempt to get a piece of the productive AI action at law these Day . While the majority of the focus stay on the poser vendors like OpenAI , Anthropic and Cohere , or the bigger companies like Microsoft , Meta , Google and Amazon , there are in fact , a lot of inauguration trying to attack the generative AI problem in a variety of way of life .
Fireworks.aiis one such startup . While miss the brand name recognition of some of these other players , it boasts the largest clear source example API with over 12,000 user , per the company . That variety of overt rootage traction tend to appeal investor tending , and the companionship has raised $ 25 million so far .
Fireworks co - founder and CEO Lin Qiao points out that her fellowship is n’t training foundation models from prick , but rather help hunky-dory melodic phrase other models to the particular needs of a business . “ It can be either off the ledge , undecided source models or the models we tune or the models our client can tune by themselves . All three sort can be served through our inference engine API , ” Qiao told TechCrunch .
Being an API , developers can plug it into their program , bring their theoretical account of pick aim on their data point , and add reproductive AI capability like inquire dubiousness very cursorily . Qiao pronounce it ’s tight , efficient and produces eminent - quality results .
Another advantage of Firework ’s coming is that it allows companionship to try out with multiple poser , something that ’s crucial in a tight - changing market . “ Our school of thought here is we need to empower users to iterate and experiment with multiple theoretical account and have in effect dick to instill their data into multiple models and test with a product , ” she said .
Perhaps even more importantly , they keep costs down by limiting the model size to between 7 billion and 13 billion parameter , compare with over 1 trillion parameters in ChatGPT4 . While that limits the universe of discourse of words the big terminology model can understand , it enables developers to focus on much smaller , focused data point sets designed to work with more special business use cases .
Qiao is unambiguously qualified to build such a system having antecedently worked at Meta , leading the AI political platform development team with a goal of building a fast , scalable ontogenesis engine to major power AI across all of Meta ’s products and services . She was capable to take this knowledge from working at Meta and create an API - ground tool that put that kind of power in reach of any companionship without requiring the level of engineering resources of a company the sizing of Meta .
Join us at TechCrunch Sessions: AI
Exhibit at TechCrunch Sessions: AI
The companionship raised $ 25 million in 2022 direct by Benchmark , with involvement from Sequoia Capital and unidentified angel investor .
