mrbonner

I wonder when Apple will separate Safari update from iOS update. They have done this for the MacOS.

I would say they their, and OoenAI, mission is to bring humanity to the next level by sharing everything they have! /s

I haven’t used Android over 10 years. But it makes it sound like iOS now, doesn’t it?

Agree. I think for the limited scopes such as in a mod/function, using Rust's borrow checker semantic makes sense as it is easier to fit in our mental model. If you have to pass those data structures around to the outside of your limited scopes, I would prefer Rc/ARc more. I also think the style guide has a catch all phrase there when it says "OK to use Rc/ARc depending on the complexity...". For me, having to pass a shared reference around is enough complexity to reason about.

I also didn’t see Ms mentions that they use Qwen as the base model. Somehow I saw a comment here and thought it was the case. Regardless, my parent comment is about other decision models that based on decoder model.

I just don’t understand the use of a decoder model to make a classifier. Jev model will give you probability scores for each of the classes/selections you ask for. The probability is actual statistical probability that a selection is right.

Using a decoder model for this comes down to picking the most probable class based on the probability what the next token assigned to a class is. They are all probabilities but semantically mean totally different things. Am I right?

“Talent” is an subjective term.

I assume Mike use the creative mode instead of survival mode. I tried survival mode and it feels too stressful for me. Even in creative mode, I struggled very much to build such elaborate landscapes a lot of folks can do. I’m wondering if this Minecraft is something you either love or hate it and nothing in between for you to learn?

What platform of Minecraft (console, pc, mobile) are you playing on?

Hah I have no idea. If we reach AGI in 10 years we will still have problems. Otherwise we all would be enlightened by then, right? Until then I think mental health would still be a problem. Maybe I should get a therapy license now.

Then I will find other problems to solve. Do you think AI will solve every problems for humanity? it has nothing to do with coding as I mentioned earlier.

What you said is valid but I think you are still constrained in thinking that the code is the solution, though. I wish I could word it a better way. Maybe we need a word calculator /s

I’ve been coding for over 20 years. I may not be the perfect coder, but for me, the real goal has always been solving problems and coding is just a side effect. If I could solve an operational problem with Excel, I would! Code, once created, is a liability I have to carry.

So I’m embracing AI-assisted coding with open arms. It’s not that I don’t enjoy coding. it’s that this approach lets me solve problems faster than ever before.

$200/month is for Navier-Stoker grade problem.

My guess is that they use an encoder-only model as the foundation and then do RLDC. Why? 1) it doesn’t need text generation 2) limited context window (40k last time I check)

Those are telltales of a Bert model.

I have been using the “free” Ling Flash model on openrouter for a bit over a month now. My work is an aside project at home building a Rust binding for an open source Zig code base library. The result is nothing impressive but also not a total failure: I have a feature parity binding to use in Rust vs. Python/java/typescript.

Now, during those night and weekend sessions, I have never run into throttling issues with the free model. Sometimes it runs a bit slow and I switch to a different free model (NVIDIA Nemo something).

So yeah, I agree with you that for professional SDE like us, we don’t consume that much tokens. I’m pretty sure the folks on the line of over limit are pure vibe coders if I can take a wild guess.

Your comment could be featured in the recent Reddit “what are the clues someone has never been poor”.

I really like this setup. Coukd you share a bit about the hardware and software in case we want to replicate? Thanks for sharing!

But that hallucination is reproducible so you can adjust the prompt. Unlike an LLM in which everything is wildly not deterministic.

It’s not only for coding. If my coworker send me a lengthy doc that obviously 100% written by AI without even reviewed/revised by them, I wouldn’t read because it is a waste of my time and an insult to me.

I mean even monkeys just dispose of the banana skin

He is not from OAI. And, did he disclose how much RSU he still keeps after “resigning” while waiting for a imminent IPO?

Does anyone use the DS Flash models for general knowledge instead of coding focus tasks? If yes, how do you rate it?

I use a 2-bit quant from Unsloth on my MBP M5 32GB of RAM. It run slower than molasses at 2 too/s kind of thing. Not sure it is usable at that rate for anything.

There is a very similar connection in the relation of skills/mcp/… and everything related to a prompt: they are all just a special-purposed prompt named differently. In a way, it reminds me of the concept in OOO: we give lots of names to “design patterns”, even write books about it. But in the end, they’re all based on polymorphism. Know this fundamental and I feel less overwhelmed.

When I was in the final year of college in 04-05 my dream would be to work at IBM. It still had the kind of prestige of R&D in the world. Fast forward a couple of years to 08-09, I don’t recall the event but IBM perception went down and a stone. Was that when they sold the hardware business and became fully consultant?

I’m interested in a general knowledge model (closed or open weight) and not coding specific. I want to plan for travel and trip. Do you have one of your favorite HN crowd?

Is it the Apple Wallet version or the actual Suica app? I'm wondering about the Apple Wallet one.

I’m going to Japan soon. Should I just use the iPhone Suica version or get the physical one?

So I think the permissive license like Apache 2 without opening up for upstream contributions would be the way of future open source development. In a very similar way of SQLite. It definitely will help shield off a bunch of AI generated contributions first and foremost and then, aligning mental model with the human contributions with Mojo leader team.

Jeez, llama.c++ is becoming the ffmpeg cargo cult CLI now