Exactly. GPT happens to hallucinate a lot of "facts" making it unreliable for a large breadth of tasks. However it's quite adept at many NLP tasks and can be fine-tuned to further improve its domain expertise.
In any sensitive application like clinical charting, one would also want to include a workflow for reviewing GPT's output for erroneous data before welcoming it in.
I'm a product designer experimenting with applications for GPT-3 in the mental healthcare space and I'm encouraged about the results of some experiments in automating the updating of a patient's chart from unstructured clinical notes.
Today, this is workflow (note->chart) is almost entirely manual consumes a large portion of clinician's time. Automating this process could free up this time to be better spent delivering care.
The fact that someone without deep NLP/NL experience can bootstrap something with this much potential impact is incredibly exciting to me.
Designer that codes. I'm a product designer by training but have moved towards engineering-focused roles (e.g. UX Engineer) to help scale GUI design and front-end development functions via in-house tooling and design system development. I'm interested in multi-disciplinary roles that combine product design and engineering responsibilities. I strongly prefer working with world-positive companies with a strong mission and values.
While React Native doesn't fully support native navigation components out of the box, there are third party libraries that interface directly with native navigation modules, e.g. https://wix.github.io/react-native-navigation. Moreover, React Native makes it easy for developers to 'drop down to' native code at any point to fully leverage native SDKs and APIs alongside their Javascript code using the NativeModules API.
It's easy to get trivial examples of native integration working, but as soon as you try more complex things it starts to break down. Code push? Oops, your native integration in old app versions just broke. Share an existing native UI library? Sorry, Yoga has its own (incompatible) thing going on. Yes, there are workarounds for all these things, but at what point does the overhead defeat the purpose?
Also, the security and performance implications of the Javascript dependency ecosystem should not be ignored.
Small greenfield app on a shoestring budget? It's perfect for that.
Although much as it pains me, I think that is a lost cause. And it was lost through attrition - devs kept throwing non-native (often, JS-based) apps at users, and users got conditioned to each app having its own look, much like websites.
I feel you. In most cases, Android and iOS's native components are superior to custom-rolled UI elements. Easier to use, more performant, and they leverage our platform muscle memory.
Fortunately, it does seem like these platforms are beginning to converge on a set of common UI patterns, e.g. Android adopting the tab bar, and iOS apps beginning to adopt Android's horizontal-paging navigation. The colors, iconography, and typography shifting can disorient somewhat, but I'd argue it's the navigation model that's most disruptive when using different apps.
It's much worse on the desktop, where we already had very detailed GUI standards, with decades of polish - and it's all getting thrown out because of the latest Electron etc fad. The upside is that Linux sees some great apps (like VS Code), but I'm not sure that's worth the cost.
We're a team of talented scientists, technologists, and clinicians on a mission to cure the most complex chronic diseases.
Our current openings: DevOps / Data Engineers / Data Scientists / Clinical Data Management Lead / Front-end engineers / Full-stack engineers / Design engineers / Mobile software engineers
Save lives with software! Docker, continuous delivery, strong team, proven founders, all good stuff.
A bit off-topic but since I got a Mac and found out about this - Google Dictionary became obsolete. The Mac dictionary works offline as well so I'm able to look things up during commutes without WiFi.
Just watched that and came away a bit confused. I'm not a minecraft player, but I understand what it is and have seen videos of people playing. I guess my question is, in the case of this, did someone go around a build that whole world just to mess with players or is it the work of a lot of different people who want to make a difficult world?
I suppose I just don't get what they are going for.
The latter. The admin takes a completely hands-off approach to running the server (I'm not even sure if he plays on it). So the condition of the spawn is purely because many people over the years thought it would be funny to make it as inhospitable as possible.
I guess it's an acquired taste, but there's something fun about trying to survive in such a hostile world. How many game servers have 200+ page comics written about them? http://minecraft2b2t.thecomicseries.com/
We should carve out a small section of land, right at the spawn. One chunk, say. Fill it with useful things like farms, farm animals, resources. Then, guard it 24/7 by using a rotating staff system. See how long we can keep it defended!
It would take a single smart person to undo hours of work. Even if you guard it really well, there's only so much you can do in minecraft.
Someone can approach you from 200m high building a 1m-wide bridge from outside your radar and just drop lava (or active TNT) on you. In minecraft, you need a pretty high defender:griefer ratio to keep things intact.
I guess it grows naturally: the first players build a huge building near spawn, some griefer destroys it all, the next builders build a bit further from spawn, etc.
Also the resources you need for building are further and further away from the spawn. Everybody takes some and decides to build just a little further away.
That video also shows the prime failure in having an anarchy server: It's so easy to cheat. The guy has a radar and can spot any player before they could spot him.
Then there's loads of other plugins that people use to find ore, hidden buildings, track players, speedboost, fly even on servers that disallow it, etc... etc...
I wanted to like Minecraft. The hilariously insecure gameplay is what breaks it. It's not a game of survival, it's a game of trolls who can exploit the most.
Our team is actually working to solve this very problem. Our goal is to make it way easier to skim and find the best parts of videos (and possibly audio later on). It's not visible on our site yet, but we have a really cool 'highlights' feature in the works that lets you quickly watch the key parts of several videos to determine if they're worth your time.
That's actually really cool. I'd be curious to see how that works with music videos, too. I also wish, however, that videos had summaries of what actually happens that we could skim ahead of time.
In any sensitive application like clinical charting, one would also want to include a workflow for reviewing GPT's output for erroneous data before welcoming it in.