This is incorrect; the local model is a small dense neural network called BirdNET that would not have been referred to "AI" when it was published in 2021. The model outputs a bird species probability distribution and the web application displays an existing image file of the most probable bird on the screen. This is a lovely example of simple, offline, and fun project using a straightforward machine learning model.
Ok, so why didn't this small dense model come up, say, 10 years back? Oh it needed all the AI evolution and concepts that were powered and evolved inside the data centers. But we still want to claim that it has nothing to do with the Big AI.
You're not wrong at all, but it's interesting that iOS launched with only web app support for third party software, and it took community pressure to persuade them to support native apps. There was not even an app store to begin with.
Pretty much every single feature added in the early days (and later) of ios was simply copying what was popular on cydia. That well started to run dry as they got better at whack a moleing jailbreaking exploits, severely contracting that community of tweak developers, and we've seen the resulting product stagnation for some time now.
I knew this guy was not worth listening to when he wrote "a solid five hours away from absolutely anything else you might want to see or do". I live in Charleston, two hours away and at the heart of the Lowcountry, one of the most historically fascinating and beautiful places in the world. And a couple of hours in the other direction are the Blue Ridge Mountains!
Charleston is an awesome city. Loved going there for work. I highly recommend Big Bad Breakfast. I believe it started there. They have this croissant-based french toast thing that is absolutely out of this world. It's 30% butter, but 130% worth it.
There's a famous paper called The Californian Ideology (1996) that shows how all these seemingly incompatible elements of the Bay Area's past created the culture at the time of the dot com boom:
Edge Impulse (https://edgeimpulse.com) is the top developer platform for edge AI: think HuggingFace but for models that process sensor data on the edge. We have a large and active community of hundreds of thousands of developers. We were recently acquired by Qualcomm, who have a crazy ambitious vision for bringing AI within reach of every embedded engineer.
This role will involve leadership, mentoring, and process design in addition to engineering, software architecture, and programming. There will be room to grow and potentially manage others as we expand the research team from 6 to 10+ researchers and engineers. Your role is to be the professional software engineer in a room full of researchers. You view writing great software as a craft, and you enjoy designing systems to facilitate quality code.
Our team literally wrote the book on AI at the Edge (from O’Reilly https://a.co/d/dIDZui0)—along with creating award-winning technologies like our FOMO constrained object detection architecture, FOMO-AD visual anomaly detection, and EON Compiler model optimization system. Our platform is cited in thousands of research papers, and we are a Gartner Cool Vendor in Edge Computing.
Like the rest of our team, you are naturally proactive and autonomous, and will go the extra mile when required: not by working late, but by going outside of your lane to help deliver the best possible product for our developers. We prize trust, collaboration, entrepreneurialism, and creativity.
I'd argue that effective written communication is a skill that can be learned and taught, and a best practice that can be enforced, rather than an innate property of a given human being.
If you can write code, you can create documents that include all the necessary details for a handover—without needing to be a brilliant writer—as long as you're aware that you're required to do so.
I couldn't agree more. Having grown up on IRC, forums, and eventually voice chat, I have no problem with "forging meaningful relationships" while working together in a text-based online world.
That said, I think it's a challenge for people who didn't exist in that space from a young age and figure out the patterns that make it manageable.
The sad thing is that these are probably easy things to learn, but there's no real movement to try and teach them. Instead, we try to replace our amazing battle-tested text-based tools with crappy multimedia substitutes.
It's definitely good to jump on a call sometimes, but I'd much rather have an in-depth technical discussion via writing. History is full of brilliant written correspondence between legendary figures in art, science, engineering, and business—but somehow we've decided that letters are not good enough for us.
The exception that proves the rule is Amazon, who apparently require written content to be submitted and read prior to in-person meetings in order to improve the discussion.
reply