Maybe I'm old but where exactly are the "dragons"?
How is RAG any different from the search systems we've been building before LLMs? Is it the sudden need for everyone to design a search API and engine that's driven this trend?
If so, I'd like to see more design patterns around existing search problems:
- Correcting or backtracking based on feedback.
- Measuring relevance.
- Comparison with task-based pre-written queries. Does every LLM task need a full blown search engine? Why not a tightly scoped domain API for data retrieval?
The whole embedding thing which converts “tokens” to vectors, which you then store in a vector database so that you can later query by vector distance, seems to be LLM specific technology, no? As far as I know the vectors look a lot like the weights in a LLM itself which is why the vector search also works with some level of intelligence.
Vector embeddings predate LLMs. They have been used as far back as the early 2000s. They are a general machine learning technique, rather than LLM specific
Unfortunately LLMs made vector search more popular so it seems like something LLM specific.
What makes it worse, a lot of people in the thread equate vector search with RAG, whereas RAG is the name for anything that model can query so a user doesn't have to copy/paste feed it to the model manually like access to text files is RAG.
Sure, the idea of making a vector embedding for words, sentences, documents etc. is old, but the meat is in how you construct this embedding. I think embeddings have gotten quite a bit better since word2vec.
It's just information retrieval through a new NN based technology that allows to map concepts and ideas as the compression of long text into points in a multidimensional space that manages to compress even more dimensions than the given ones, through non-transparent engines that give different mappings and results, and still (the information retrieval) requires many more clever tricks than the simple idea of vector distance ordering because things do not quite work as they should.
Let's say it's just "computation packaged as something new". "Trivial things".
> The idea of reactivity came from or was popularized by Angular.js
No, spreadsheets popularized reactivity. And the general point is incredibly weak.
Don't use frameworks and make your own? Sure, have your fun. But then try teaching your framework to your company of 1000 and see how quickly you realize your view of the "problems" are only a slice of the pie.
Aside from the Giants maybe some consultancies or hire-firms have that many? But they use whatever the client uses. If you're not in the business of selling frontend development services then a big part of your competitive advantage is _being able_ to make technology choices that are more-optimal than the labor market's lowest-common-denominator slurry.
Between JS/ECMAscript being abandonware for a decade and CSS' historical immaturity (cross-browser layout, custom properties, more pseudoclasses, etc) the Web was more of a wild west "whatever you we can cobble together" place. So of course we built an entire toolchain ecosystem for all the cool new tools we would reinvent from scratch.
but Things Are Pretty Good these days. We even have the world's only fully system portable assembly language: WASM! And the browser's Web APIs are becoming something like the missing Standard Library (shoutout to Deno).
The author's point seems to be that we are past a tipping point where we need all that additional complexity. You don't _need_ a separate system of reactivity these days, you can write a (closer to pre-hooks React, even!) React-style component library with Web Components now.
Personally I would need for Type Annotations to be a fully supported part of the spec before I made the point to "use the platform" professionally. Until some market force causes the core web technologies to truly fork my skills are useful for life. It's like a pale shadow of how UNIX users must feel :')
My point was your needs and problems are not everyone's needs and problems.
I don't want to think of memory management in most projects, I want to focus on business logic, so I use Node.js instead of Rust, even if the latter gives me more control.
Similarly, I don't want to think of DOM management in most projects, I want to focus on data dependencies, so I use React instead of Web Components, even if the latter gives me more control.
Personal and professional are not mutually exclusive.
If I criticize your code, that is a professional criticism.
If I criticize your code and say it reflects your consistent carelessness and stupidity, it is also personal.
If I say you fabricated something, then that is a personal criticism, it alleges an ethical violation. In a professional context, it's also a professional criticism (every profession has some ethical standards).
Can you criticize a project which is mainly contributed and managed by one person without criticizing the same person who does the decisions that cause criticisms?
Yes, I think very easily and I have read examples of this. There are bits of this in the article, but the main thrust attempt to portray Jarred as a greedy asshole enamored with Thiel/VC thought is not about the project and quite clear reading the article. It’s entirely tactless and bitter imo
> main thrust attempt to portray Jarred as a greedy asshole enamored with Thiel/VC thought
What made you get that takeaway from the article? I didn't get that feeling at all, mainly seems to be something like "Jarred does some good and some bad, personally I don't agree, still wish him well", but clearly some specific part in the article must have given you this impression, if so what part?
The "some good" part reads like it exists as a buffer between other parts which don't sound objective, but rather defensive and personally angry.
> Fun fact: people talk to each other.
The intention here seems to indicate that Jared could've never known that could happen. It doesn't sound like professional feedback and more how you talk to someone during after a road rage incident.
And the fact that immediately after no "personal criticism" he proceeds to call his behavior "fantasy fever dream." Sometimes presentation matters as much as factuality.
The lack of sources and citation for a pretty one-sided claim doesn't help either.
> jumping head first into problems that he was not yet equipped to solve, leading to mediocre outcomes in terms of engineering
> having graduated from the Thiel Fellowship school of thought rather than university, he was essentially groomed from a young age into uncritically embracing the Silicon Valley mindset,
> this "beginner energy" started to hit differently for me. It's one thing to choose a poor work-life balance for oneself; a different thing entirely to demand it of others
> Poor communication, unrealistic expectations, low empathy, no experience
> gets to live out his productivity fantasy fever dream, he's probably already super wealthy. He has minor tech celebrity status.
Frankly, I have trouble seeing how a neutral reader doesn’t see this as a clear personal attack. That the article ends with “I actually don't have any personal criticisms of Jarred” is almost comical given the preceding paragraphs.
> The Zig foundation had a continuing disagreement about how Bun was using Zig (their methods and the resulting code). Other projects that follow the advices of the Zig foundation more closely won't have the same problems Bun had.
Posting this could've been enough to save face.
I find the blogpost super petty, 8-10 untrue, unprovable, unrelated jabs against a person and colleague.
And literally 3 sentences later he goes back to insulting him ("productivity fantasy fever dream"). Even if that is true, it's still an unwise post to publish in this form IMHO. If the goal was to defend Zig, that could've been done in a less personal manner.
I think he messed that part up and it comes off as passive aggressive but he is probably scared of the “outrage” from the angels of the internet about how rude he was to Jarred
> Jarred was already writing slop well before he had access to LLMs.
I don't see anything business related in that statement.
What's this new level of gaslighting? "It was not because of me, but because of the business situation I was in". Wait... wasn't he in that "business situation" because of actions HE took?
Not prevent, just not provide very responsive feedback, right?
I don't know, I understand the principle, but I don't see how you can determine the value of a principle outside of a specific context.
Even for accessibility, we can't target every context in the name of being accessible. We still have to pick which contexts of inaccessibility we'll need to support with more attention.
Lots of people assume that the valuable thing is the direct business, but a business can be a lot more than that. A competitor may buy you for your engineers or sales team or patents, or assets, or whatever (e.g. Siri, Motorola, etc)... and just toss the rest of the business or sell that stuff off after they have what they want. In other industries they may just buy companies for their assets. Also you never know what will happen with a pivot (e.g. twitter, slack, etc).
The bleak reality is if you can keep growing (making more money than you spend) that alone is usually desirable enough for people to keep giving you money and eventually provide an exit. Why? because it's really hard to do. Its a skill.
And I have to say that no one tries to build a failed business. Founders can be really earnest about their intentions, work harder when they see the cracks, but often it just doesn't work, they don't find the right way before it's too late.
Maybe just before the end someone tries to siphon the funds into private account or assets into their next venture, but you tend to get caught doing so.
How is RAG any different from the search systems we've been building before LLMs? Is it the sudden need for everyone to design a search API and engine that's driven this trend?
If so, I'd like to see more design patterns around existing search problems:
- Correcting or backtracking based on feedback.
- Measuring relevance.
- Comparison with task-based pre-written queries. Does every LLM task need a full blown search engine? Why not a tightly scoped domain API for data retrieval?