chrischen

Actually I found the coding models to work really well with these languages. And the type systems are not actually complex. Ocaml's type system is actually really simple, which is probably why the compiler can be so fast. Even back in the "beta" days of Copilot, despite being marketed as Python only, I found it worked for Ocaml syntax and worked just as well.

The coding models work really well with esoteric syntaxes so if the biggest hurdle to adoption of haskell was syntax, that's definitely less of a hurdle now.

> Instead, you're writing non-trivial proofs about all possible runs of the program.

All possible runs of a program is exactly what HM type systems type check for. This fed into the coding model automatically iterates until it finds a solution that doesn't violate any possible run of the program.

We already have verification layers: high level strictly typed languages like Haskell, Ocaml, Rescript/Melange (js ecosystem), purescript (js), elm, gleam (erlang), f# (for .net ecosystem).

These aren’t just strict type systems but the language allows for algebraic data types, nominal types, etc, which allow for encoding higher level types enforced by the language compiler.

The AI essentially becomes a glorified blank filler filling in the blanks. Basic syntax errors or type errors, while common, are automatically caught by the compiler as part of the vibe coding feedback loop.

I would guess HN readers are not an average cross-section of broader society, but I would also guess that because of that HN readers would be pretty bad at understanding what broader society is thinking.

I guess the translation can always update itself in real time if the model is fast enough.

Some of your points are already considered with current implementations. Airpods live translate uses your phone to display what you say to the target person, and the target person's speech is played to your airpods. I think the main issue is that there is a massive delay and apple's translation models are inferior to ChatGPT. The other thing is the airpods don't really add much. It works the same as if you had the translation app open and both people are talking to it.

Aircaps demos show it to be pretty fast and almost real time. Meta's live captioning works really fast and is supposed to be able to pick out who is talking in a noisy environment by having you look at the person.

I think most of your issues are just a matter of the models improving themselves and running faster. I've found translations tend to not be out of whack, but this is something that can't really be solved except by having better translation models. In the case of Airpods live translate the app will show both people's text.

This is the sad reality of most if these AI products and it’s that they are just taking poor feature implementations on the hardware. It seems like if they just picked one or these features and doing it well will make the glasses useful.

Meta has a model just for isolating speech in noisy environments (the “live captioning feature”) and it seems that’s also the main feature of the Aircaps glasses. Translation is a relatively solved problem. The issue is isolating the conversation.

I’ve found meta is pretty good about not overdelivering on promised features, and as a result even though they probably have the best hardware and software stack of any glasses, the stuff you can do with the Rayban displays are extremely limited.

Real time translations is a real good use case. The problem is most implementations such as the Airpods live translate are not great.

Current LLMs are doing this for coding, and it's very effective. It delegates to tool calls, but a specialized model can just be thought of as another tool. The LLM can be weak in some stuff handled by simple shell scripts or utilities, but strong in knowing what scripts/commands to call. For example, doing math via the model natively may be inaccurate, but the model may know to write the code to do math. An LLM can automate a higher level of abstraction, in the same way a manager or CEO might delegate tasks to specialists.

But these models are more like generalists no? Couldn’t they simply be hooked up to more specialized models and just defer to them the way coding agents now use tools to assist?

I think a bigger problem is the HN reader mind reading what the rest of the world wants. At least when an HN reader telling us what they want it's a primary source, but reading a comment about an HN reader postulating what the rest of the world wants is simply more noisy than an unrepresentative sample of what the world may want.

Honest companies are priced out by scammy companies, and as long as these companies share the profits they are totally fine profiting off scams. They make more money off the scams, simply put.

Agreed. At best most of the stuff I ended up buying from an Instagram ad turned out to be oversold or overpromised and underdelivered. While not a scam outright, it's sort of training me to avoid buying anything from ads...

No need to abuse anything until shit hits the fan.

Maybe what we need is AI based code review.

Not sure how many people are playing real tennis these days. But for most older established sports there are existing communities and apps that service them which make it hard to penetrate. Whether it's tennis or real tennis, even with badminton I have a hard time convincing groups to use it. The software is provided as-is so if people use it then that's fine, but otherwise I'd mainly target the ones I actively organize around so that there is actual usage. There's also not much in it specific to any single sport so supporting real tennis, however obscure it is, is basically free from the software side.

Is it that flawed? Maybe a bit early and not enough cash behind them as say a company like Meta or Apple (planning to pivot the VR headset into AR glasses).

Why allow the sale of a nationally important company if they weren't willing to let it go? Seems like eating your cake and having it too.

Made a partially open source event platform for Pickleball and Badminton in Asia. Apart from managing event attendance, reminders, waitlist, level restrictions, and calendar syncing, there is also a pretty thorough Event Manager built in.

The event manager has a built-in ELO-style rating system designed for computer games and applies it to offline sports. You tap on the winning pair and it automatically calculates new ratings offline. It takes the attendees from the event automatically, and can manage taking turns evenly, placing players of similar level together and balancing matches. It can also do all of that while handling players who come in and out at random times, or matchmaking for fixed pairs.

So what problem does it solve? The event manager is a solution for anyone who organizes slightly larger groups of players. So instead of a a paddle queue or random-round robin tables, an event can be managed by level at a very granular level fully automatically. Players just go to the court they are assigned and they'll be put in a balanced competitive match automatically. At the end of the match the players just tap on the winning team. I think commercial facilities or clubs could also benefit from using this as it can do matchmaking for "open play" style events, but this is an open source side project and I have no intention of making any sales pitches.

https://www.pkuru.com - please check it out and if you are visiting Tokyo or Bangkok you can join the Pickleball events straight from the website.

Another way to look at it is what's the real damage considering the breach right under Vietnam airlines already leaked that.

I've never understood monthly subscriptions. It's not like it's technically difficult to implement prorating or even look back and see that I haven't used Disney Plus at all and refund me for non-usage.

This event probably triggered a lot of their inactive subscribers like me to look at their bills. Not only will it lose a lot of these people but the required marketing efforts to bring me back would cost a lot more now too.

To be fair, I think a lot of people are in fact using it like a slot machine, which is where a lot of the "AI doesn't help me code" perspectives are coming from.

When I say ideas i'm talking in the context of programming... I'm not talking about "I got a great idea for a new social network" and the AI just wrote some spaghetti code for it. When I have the AI write low level code it's stuff like filling out a function implementation of which I already defined high level type classes for... I can focus on high level abstractions, whereas the AI can iterate in the most statistically sensible way to fill in the easy blanks.

Exactly. I loved doing novel implementations or abstractions… and the AI excels at the part where it modifies it slightly for different contexts… aka the boring stuff.

Walling in my personal data would be a sure sign way to get me to not use OpenAI.

I think you can make arguments for and against the fundamental right to record or to not be recorded.

If someone is doing something bad/illegal, do we have a right to record/document it? If I am outside minding my own business and not doing anything bad, do I have a right to not be recorded?

What is the difference between seeing and recalling something that happened vs recording? What happens when technology blurs the difference (for example if we all start wearing and using camera AR glasses)?

I don't understand why people take bad coding practices and just let AI run with it and then expect nothing but poor quality code. Nothing about the AI revolution here changes how good software has always been written. Write tests, use a typed language, review code. If you have good patterns, good procedures, AI fits right in and fills in the blanks perfectly. Poor AI results tend to be the pot calling the kettle black.

The just released a couple of styles with the announcement of the new ones, but before that the low bridge fit ones got discontinued. But you are limited to a specific model/frame.

There’s a Chinese AI glasses brands (Solos) that integrates with ChatGPT and I’m wondering if simply being paired with a better AI model will make it significantly more useful. One thing Meta seems to be completely ignoring up until now is Asians and Asian markets (lack of low bridge fit models, lack of translation features for any Asian languages despite ChatGPT being state of the art at it).

That's probably due to general inflation...

HTML is an implementation detail and JSX is a syntax for describing UI components. For example JSX can be reused across different contexts such as for React Mobile. It just so happens most usage of JSX is used to render HTML so it superficially appears to be writing HTML syntax. But I prefer JSX because it allows me to write UI HTML components as pure functions, without having to syntactically write out actual functions in whatever programming language I'm using.