They state further down that they're working on non-Nvidia support. Looking forward to it, since I'm pretty heavily invested in suffering on AMD (ROCm sucks, but everything else about AMD is worth it to me.).
I have noticed 5.3 in xtra high was a turd today. High used to be enough for most of my use cases. xhigh used to surprise me. Now it's incapable of following the very first instructions.
I just hope open source models get as good as last few month's top models before the enshittification has gone too far.
Exactly. Open models are a wrench in monetization plans. If a free model exists, and it’s ad free, then why pay for the proprietary model that comes with ads? It’s a worse product! Presumably you’d just use the better experience at that point
I built a jira with attachments and all sorts of bells and whistles. Purrs like a kitten. Saas are going extinct. At least the jobs that charged $1000 a day to write jira plugins.
Some minor UX enhancement SaaS of the most recent VC-funded wave will do. Maybe those who forgot how to invest in R&D and spent last 20 years just fixing bugs. There’s plenty of SaaS on the market that offers added value beyond the code. Data brokers. Domain experts, etc. Even if homemade solution is sometimes possible, initial development costs are going to be just one of several important factors in choosing whether to build or to buy.
SaaS are not going exctinct. This reminds me of the LinkedIn posts saying they clone Slack in two hours, copying the UI, etc. Yeah, if you think Slack is private chat rooms then you should use IRC for your company.
One of the most valuable things about Slack is the ecosystem: apps, API support, etc. If you need to receive notifications from external apps (like PageDuty or Incident.io or something like that), good luck expecting them having a setup for your own version of the app. Yeah, some of them provide webhooks (not all of them), but in the end you have to maintain that too...
Yes, it seems like it got to some tipping point around 2013 where so many product and management people were familiar with it, and from there it became this “industry standard” that management always wanted everyone to use.
Also though, I feel like being attached to Confluence helped it because there is a lot less competition in the world of documentation wikis than there is in task management.
In Public mode, Ceno will look into the BitTorrent network to see if another Ceno user has recently shared the requested page. If the service can identify the requested page, it will retrieve that page from another user's device. If the content is not available, Ceno will contact several Injectors to request that website and have it delivered to you.
In Personal mode, you will only contact the Injectors to have that website fetched and delivered to you. The search will not connect to the BitTorrent network and will not attempt to locate the content on other users' devices.
To ensure that your Ceno client can always contact an Injector, we have also created Bridges. If the Injectors are blocked on your network, the Ceno app will look for available Bridges, who will forward your request to the Injectors. The Ceno network currently features around 6,000 Bridges. Their number is always growing.
So on the one side it's some kind of shared cache of website resources, and on the other some kind of distributed tor-like edge network?
Quite clever! I wonder if it works well though, and if there is a risk of content injection by adversaries.
If anything it shows the applicant was over qualified and it probably wasn't a great fit; to have this guy, who has battle experience, in a place where the interviewer was looking for a one-size-fit-all canned type answer.
Is there an alternative, tutorial, or project you'd recommend that would help me do supervised fine tuning (SFT) with the metal stack / macOS?
reply