what's happening at askabl:

July 2023

  • July 22nd: Main video: The main video is up. We are still debating whether we should use the more modern H.265 codec (aka HEVC), or stick to the older H.264 standard, which should be compatible with virtually any device out there visiting our site. Probably our best bet is the latter, even if there is a lot to like about HEVC: it provides the arguably best quality to filesize ratio out of all video codecs. But there may be people on devices visiting our website which do not have a dedicated (hardware) HEVC-decoder, or a processor powerful enough (although the part of those devices is small, and will diminish even more over time). Hence, we're going with the safe option and use the standard that works for everyone - plus, we're talking about a mere few MB's of difference in size for the main video encoded with these different standards, so there is no real reason not to go with the option providing the widest compatiblity.

  • July 12th: Website: The first draft of the technicals page is up. It is incomplete, and the first passage is way too lengthy ... but all of that will be fixed in time. We are just happy, that technically speaking all of the essential pages are up now, and we will be in the process of improving and adding stuff over the next period of time. Naturally, the main page will need the most work, but there are also pages like the ethics page, which is already sort of ok, in our opinion. So the point where askabl's website will launch is creeping up.

  • July 12th: Website: Although the website's main page is not finished, the ethics page is, a blog is easy to do ;-), the technicals are still absent, but work continues on all of them. Most notably missing from the main page is the main video, as well as development stages / roadmap. In any case: the askabl website will launch soon (which means simply removing the currrent "work in progress" landing page, instead redirecting traffic to the then complete main page, and remove our current robots.txt, which prohibits all search engines from indexing). We are all very excited ... and curious, what feedback we might get. Supposedly the site will get little traffic in the beginning, and even if someone stumbles onto it, he/she will in all likelihood not quite believe it - or maybe not at all. We remain confident, that once people get to take askabl for a test drive, they will come around.

  • Jul 11th: Test runs: Those will naturally be done on a powerful server. Already looking forward to testing askabl on varying hardware. There is a lot to like about Apple's approach of designing very energy-efficient, yet powerful chips - M-chips are astonishingly far ahead of the rest of the industry when one compares wattage to computing power. For this reason, they are also the most ecological option, and no other CPU/GPU/SoC comes even close. But askabl works better on an x86 platforms, due to a couple of technical reasons. Still, there will be an M-chip optimized version of it some day - when we started the project, we could hardly foresee this development, so we coded/optimized for x86. Also, we still have to evaluate if (and how much) some parts of the code can be run faster or more efficiently on a GPU. Currently, we are simply trying to get done something like a "minimal viable product" (although askabl is not really compatible with this development strategy). askabl needs a well-coded base: this is essential to the whole project. Thus, we cannot just push something half-baked, cannot sprint ahead for an mvp, and instead are focusing on a solid basis first, while some components may be done the mvp-way. As is explained on the main page, and even more so in the (forthcoming) technicals page, askabl needs much more coding work than any other NLP, or AI project at all, we are aware of. So everybody better be patient, including us ;-)

  • July 8th: Work: Usually, the individuals involved in askabl all have their focus, their specialty within the project. Yet, we have decided to pool our resources for now, so that the website gets done quickly. This naturally draws time and focus away from software/backend dev, but a) designing a website is fun, and b) collaboration is always better. Hence, we are focusing our efforts now, together - and hope to be able to be able to attainx a presentable site by the end of the month.

  • July 6th: Ethics page: The first version of the ethics-page is up. This is still a first version, but it is likable the way it is now. Simple, but good. It is actually the page that is most done of all - if one can even talk about a website being done, since it is not supposed to stay static. Arguably, however, an ethics/mission statement should stay the same, so maybe this is the final version of the page :-)

  • July 4th: Header: Updated the header image for better readability. Still not good enough, since on screens with a strong horizontal ratio, i.e. wide screens, the image gets cropped too much.