I don’t think you should have to pay extra for extra security in general. Making a product or service free of security defects ought to be considered a basic requirement of merchantability.
But we should also draw a distinction between, like, real security defects (RCEs, that sort of thing) and features that might make it easier to deploy a system securely (SSO).
... Because the specifications are open. Practically the whole Internet is built on open specifications. The security and operations benefits are obvious for enterprise customers. Startups could also benefit greatly from it, but the cost ramping of the large providers is onerous.
Hi there! Thank you for the glowing review! I'm the cofounder of Krea and I'm glad you liked Sangwu's blog post. The team is reading it.
You'll probably get a lot of replies around how this model is a just a fine-tune and a potential disregard for LoRAs, as if we didn't know about them. While the reality is that we have thousands of them running in our platform. Sadly there's simply so much a LoRA and a fine-tune can do before you run into issues that can't be solved until you apply more advanced techniques such as curated post-training runs (including reinforcement learning-based techniques such as Diffusion-PPO[1]), or even large-scale pre-training.
definitely; our main goal is to see how much we can accomplish in a month and do a mini social demo day at the end. more info in our faq https://www.hackerresidencygroup.com/faq
krea.ai | Distributed Systems Eng. & Data Engineers | San Francisco (on-site), CA | $200K-$400K
I'm the cofounder of Krea (https://www.krea.ai), a startup in San Francisco building browser-based AI creative tools to improve the workflows of artists, designers, and creatives. Small team (~15); millions of active users; and we've raised over $83M from top silicon valley firms.
We're looking for talented engineers who want to tackle hard technical challenges with smart people while building a creative platform that big companies and startups will rely on.
The job:
- Build distributed systems to process massive (billions of files, petabytes), amounts of image, video, and 3D data, solving scaling bottlenecks as you go.
- Work with our research team to build ML pipelines and deploy models to make sense of raw data (example - find clean scenes within videos).
- Learn ML engineering from world-class researchers on a small tight-knit team. As you’re building data pipelines, you’ll contribute to foundation image, video, and world models from 0 → production, and see them used by millions of Krea users.
- Play with massive amounts of compute on huge Kubernetes GPU clusters; our main GPU cluster takes up an entire datacenter from our provider.
Our tech stack: python, pytorch, k8s, a rotating cast of data tools (e.g. DuckDB, massive relational dbs, PyArrow, etc.)
You should apply if you are an excellent generalist engineer with strong backend experience and an intuition for systems design. Bonus points for deep experience with distributed systems, kubernetes, and existing ML or data experience (though not required!). Cool side projects are a green flag.
Can you point to a URL with the tests you’ve done?
Also, FWIW, this model focus was around aesthetics rather than strict prompt adherence. Not to excuse the bad samples, but to emphasize what was one of the research goals.
It’s a thorny trade-off, but an important one if one wants to get rid of what’s sometimes known as “the flux look”.
Re: Wan 2.2 I’ve also been reading of people commenting about using Wan 2.2 for base generation and Krea for the refiner pass which I thought was interesting.
The Image Showdown site actually does have Flux Krea images but they're hidden by default. If you open up the "Customize Models" dialog you can compare them against other Flux models (Flux.1 Dev and Kontext).
> FWIW, this model focus was around aesthetics
Agreed - whereas these tests are really focused on various GenAI image models ability to follow complicated prompts and are not as concerned with overall visual fidelity.
Regarding the "flux look" I'd be interested to see if Krea addresses both the waxy skin look AND the omnipresent shallow depth of field.
reply