If you’ve been building features with LLMs over the last few years you’ve likely used frameworks like LangChain/LangGraph, Vercel AI SDK, LlamaIndex, Google ADK, etc. Alternatively, you may be deeply exposed to chatbots, coding agents, and/or assistants. All of the above, roughly circling around the same idea, are flavors of the AI Agent loop being a mix of large language models and software scaffolding (typically called a harness).
What’s surprising is that for Swift and Apple platform developers, a solid version of the harness-building toolset doesn’t seem to exist yet. Working on some exploratory features, I kept having to relearn and rebuild the same scaffolding every time I wanted to try a different provider. The kind of thing a framework exists to solve. And since a clean Swift abstraction of it didn’t already exist, I built mine.
While I flattened the repository history before making it public, the repository history of AgentKitten spanned more than 570 commits and 130 pull requests of rethinking the shape of things. And I've continued to maintain it since, having piled up 64 more PRs to improve it further. It is built using AI coding tools, but steered carefully and deliberately.
I believe we still have a world of opportunities to put AI (particularly small on device models) to
the service of improving mundane interactions in our apps (I put some samples in the repo’s README.md).
I think this focus was kind of instrumental on how I built AgentKitten. It is more powerful that I expected it to
be when I started it, but still I tried to stay grounded about being practical for the little use-case.
Clone it and run `swift run Playground --help` to see examples of what it can do.
Without widespread adoption is still an early stage project and I'm looking for constructive feedback to improve it.
I don't think that they even expect them to be predictors of importance or popularity. The "best paper" award can give a hint of temporary fashions or about what seemed to be interesting concepts/works at a certain point in time. That's why it seems especially interesting to see them compared with the "top cited" in these charts.
I'm nowhere near to be techno-apocalyptic, but I just thought that from a "1984" point of view it could be a tool to monitor activities without no permission at all.
There are many things I'd expect as such tools/customisations... A pair of simple examples:
* Specialised support: the croudfunding platform could offer legal and strategic support in the narrow field they are focusing. I think Kickstarter cannot embrace all industries even if they try.
* Specialised tools and rules: Each platform must have extremely distinct interactions with the backers and product outcomes. In an academic platform (as the one being discussed) I'd expect some tools for scientific divulgation (which must be a mandatory "return" from crowdsourced research projects) and collaboration (which must be extraordinarily stimulated). I think Kickstarter will never consider/establish these rules/tools.
Don't you think that specific industries may need specific tools that Kickstarter doesn't provide?
And the "return" is absolutely different...
I think that these simple things are the sparkle to create the need for separated platforms.
However, many of these alternative croudfunding platforms are not implementing any difference. And there I will support your point of view, there is no need to do that. In fact, it should be better to take advantage of Kickstarter's moment of inertia.
Having to depend on Universities or any other institution for security's sake seems to be a little bit restrictive. Individual or non-institutionalized research should be acceptable. Responsibility with project backers shouldn't depend on that, don't you think? There should be a legal contract anyway.
However, I don't think that this will be the most common profile looking for funding.
I'm not way ahead of you (currently I'm finishing my PhD), but I'll try to help you because at some point I had the same question.
I sticked to the CS degree. Let me argue that by answering some of your specific reasons:
1a) Grading system is kind of broken, it is absolutely true. It is too tempting for students to focus on grades rather than on learning. But that is until you realise that. Now you can do both! It is good you've noticed it before your degree ends. From now onwards, you'll be able to combine learning and passing exams. In fact, I you do so, while learning, rules to pass the exam will become easier to understand.
2a) They may o may not be teaching practical stuff, but remember that practical stuff usually has a shot live span. Learning different stuff at the degree will exercise you to be able to adapt to any future stuff. Besides, you should be able to appreciate all non-academic things that you learn while you are at the faculty.
1b) The world shouldn't be like this, but maybe in the future some people will judge (and pay) you by your résumé rather than your abilities… but, whether you get the degree or not I'd urge you to run away from that kind of people. There will be more opportunities in CS market.
3b) One of the many evaluations of a degree is that you must show you can commit to all adversities (grading system, professors, team work, etc.). You must prove that you know how to get to the end of any journey you've begun. Many people will look for that. However, the world is also full of happy successful academy dropouts and people who never get to it from the beginning. You can prove your value without academia.
IMHO this numbers must be compared with employment rates of people who haven't got into a PhD. We may discover a similar (or even more significant) downfall. I think that it is not an awful market specifically for young scientists, its an awful market for everybody.
Sometimes even the topic is what makes the difference (not even location or device) between talking synchronous or asynchronously. I'm not a very assiduous Facebook user, but at some point I've noticed it can blur the edge between sync/async. Maybe it is not perfect (and maybe it is no longer there), but I think its possible to mix both ideas. Neither should be left out.
Wow, I didn't know they charged 99 cents/year to Android users now. It wasn't like that in the past! iPhone monetization continues as a single payment. Do you think that this is enough to maintain the company? Many other companies should be perfectly capable to reproduce the contacts-promiscuity that Whatsapp is using. And they could even mix it with compatibility to other XMPP platforms...
I never fully understood why everybody is using XMPP in walled gardens (e.g. Whatsapp, Facebook, Messages...). Do you think they get anything from that? Is any of those services monetized successfully?
Indeed, mastering Vim is a result of practice and not of configurations and plugins. However, most people get sucked into Vim to have this personalization, and only a few stay to really master Vim. In other words, almost everybody goes through that step, hence the "configuration madness". I don't think it is inherently bad, it is part of Vim's wizardry.
It is not being reindexed (it is really difficult to know when it is actually triggered). Do you think that it is a problem? You could let them notice this flaw at http://www.apple.com/feedback/
What’s surprising is that for Swift and Apple platform developers, a solid version of the harness-building toolset doesn’t seem to exist yet. Working on some exploratory features, I kept having to relearn and rebuild the same scaffolding every time I wanted to try a different provider. The kind of thing a framework exists to solve. And since a clean Swift abstraction of it didn’t already exist, I built mine.
While I flattened the repository history before making it public, the repository history of AgentKitten spanned more than 570 commits and 130 pull requests of rethinking the shape of things. And I've continued to maintain it since, having piled up 64 more PRs to improve it further. It is built using AI coding tools, but steered carefully and deliberately.
I believe we still have a world of opportunities to put AI (particularly small on device models) to the service of improving mundane interactions in our apps (I put some samples in the repo’s README.md). I think this focus was kind of instrumental on how I built AgentKitten. It is more powerful that I expected it to be when I started it, but still I tried to stay grounded about being practical for the little use-case.
Clone it and run `swift run Playground --help` to see examples of what it can do.
Without widespread adoption is still an early stage project and I'm looking for constructive feedback to improve it.