It seems it's only based on the transcript for now. For example when I asked to comment on visual content it mentioned requiring a captions file:
> I'm sorry, but I'm unable to access this YouTube content. This is possible for a number of reasons, but the most common are: the content isn't a valid YouTube link, potentially unsafe content, or the content does not have a captions file that I can read.
No. It just hallucinated[0] about parallel evaluation while this talk doesn't mention parallel evaluation at all. So it has to be more than just the transcript.
It was updated just now with a message saying "Users are having trouble connecting to Slack". So about 15 minutes since the comment above. However, I think the first reports started a bit earlier, so maybe 30 minutes is a better estimate.
Very excited to see modular and repairable tech in production! Just wanted to make a top-level comment in support of a trackpoint option. Given modular design and the Marketplace it could make sense to have a simple drop-in replacement for the keyboard module that would potentially lure many hardcore Thinkpad fans.
It may be a nice feature, but for me Firefox's implementation seems to be a bit too intrusive, because as a developer I can't disable it and it is always visible as a browser-styled overlay.
Not all video elements are for stand-alone consumption. For example, there are "hero" videos in landing pages, enterprise solutions requiring smart overlays, or imagine a component in an online video editor.
Design is iterative by its nature – therefore FROONT can be used for prototypes and for designs, that depends of the stage and complexity of design. IT can be wireframing, designing responsive patterns or design with real content.
During the process of developing the tool, it has evolved from a prototyping tool into a design tool. Unfortunate delay in updating the definition on CrunchBase.
This is exactly what we are doing in toostis.com. We try to create a one-in-one tool for sharing this kind of information with your interest circles (that we call "tribes"). Currently when you log in and start attending events the recommendation algorithm learns about different social groups you belong to.
One needs to consider three approaches of forming an interest circle: shared (an explicit list of members), individual (each user maintains several lists of his friends) and automatic (extracting subgraphs of user relation according to attendance).
What we are currently willing to provide is the combination of them all. Automatic extraction is fun to play with, but you always want more control over it. We are also preparing to launch something similar to the shared model, where you could subscribe to an organizer which is basically a group of related events.
> I'm sorry, but I'm unable to access this YouTube content. This is possible for a number of reasons, but the most common are: the content isn't a valid YouTube link, potentially unsafe content, or the content does not have a captions file that I can read.