[ad_1]
Replication, a startup that runs machine learning models in the cloud, has floated today with $17.8 million in venture capital backing. Of the total, $12.5 million came from a Series A led by Andreessen Horowitz at Y Combinator, Sequoia and Angel Investors including Faygma CEO Dylan Field and Vercel’s Guillermo Rauch, and the remainder from a previously undisclosed seed round.
The company was co-founded by Ben Frishman, who leads open source product efforts at Docker, and Andreas Janssen, formerly a machine learning engineer at Spotify. In the way Frishman tells it, he and Jansson came to the common understanding that AI is accelerating at an “absurd” rate, but that technical hurdles stand in the way of mass adoption.
Enter Replicate, which provides a library of open source models that software developers can work with with just a few lines of code. The platform can automatically generate an API server for custom machine learning models deployed on a large set of GPUs.
“If you get a ton of traffic, we scale to meet the demand. If you don’t get any traffic, we drop to zero and pay nothing,” Frishman explained. We only pay you for how long your code runs. The alternative is often to deploy models on Amazon Web Services. Typically, from servers, You have to contend with Kubernetes, GPUs, API servers, auto-scaling, and more.
Core to replicate Cog is an open source tool that allows developers to package machine learning models in a standard, production-ready container format. Frishman and Janssen created Cog, which runs on any new macOS, Linux or Windows 11 machine.
“AI is very difficult for software engineers to use these days and you have to be a machine learning engineer to use it,” said Freshman. “Companies and the industry as a whole have been held back by a lack of machine learning experts. We’re enabling software engineers to use machine learning with zero experience, with just a few lines of code, to build products with AI and apply it to business problems.”
Diffusion supports thousands of ready-to-use models, including text-to-image and image-to-text models (à la Stable Diffusion). Image Credits: Repeat
It’s not just replication that does this. The startup competes with vendors Hugging Face and OctoML (and to a lesser extent Runway ML), which have collectively raised hundreds of millions in venture capital. Google, Amazon and Microsoft can be considered rivals – they offer their own solutions for developing, launching and maintaining machine-lean models in the cloud. (See SageMaker, AutoML, and Azure’s no-code ML tools).
So what is different about multiplication? Freshman says the developer experience is “much better,” which really remains to be seen — after all, Replicate is so new. One clear point of difference, however, is the expansion of the Replicate AI library. The platform offers stable diffusion models for creating and editing videos, enhanced models for images, and a variety of image-to-text and text-to-image models.
Fast, painless deployment is the focus. Replicate’s website promises: “With Replicate and tools like Next.js and Vercel, you can wake up with an idea and watch it hit the front page of Hacker News until you go to bed.
The deal seems to be resonating with the developer community, which has enthusiastically embraced Replicate over the past few months—at least according to Firshman. The platform has seen 149% month-on-month growth since the middle of last year, he said. Enterprise clients include Character.ai, Labelbox and Unsplash.
“We were effectively pointing to the progress in generative AI,” said Fershman. “Founders are building a lot of new products, investors are investing in it and users are clamoring for all these new things.”
Leaning towards generative AI is definitely a wise decision on Replicate’s part. The segment – where technologies like ChatGPT and Stable Diffusion fall – has seen a significant increase in investment over the past several years. PitchBook (via Bezinga) reports that VCs have invested 425% more dollars into generative AI in 2022 compared to 2020, with the space reaching $2.1 billion in total capital pledged by 2022.
Freshman continues to grow – and replicate the benefits.
“It has not yet entered the consciousness of the enterprise how creative AI will enhance many of their business activities: customer support, marketing, sales, content creation and perhaps other things that we have not yet anticipated,” he said. “In the near future, customer support will be largely automated and highly efficient – not the dreaded chatbots of the past. Creating marketable assets is largely automated. Most of the ads you see will be automatically generated and personalized. Creating assets for video games is mostly automated. And this is with the technology we have today.
[ad_2]
Source link
