s16h

Oh, and this was entirely vibed-coded by https://x.com/probabletrain — wild.

riffrail is a browser-based music sequencer and synthesizer. create beats and melodies by connecting nodes together, then hit play to hear your creation.

Oh, that’s me. Just worth noting—the picture was taken as a joke, but they ended up using it anyway.

Not that this is the whole solution, but I can send you a used MacBook Pro M1 or M2 (can’t remember). You can use that to go on Mercor and see if any of the roles suit you.

Reply to this thread, if interested.

I had fun vibecoding and fine-tuning a model on all past proberts answers: https://news.ycombinator.com/item?id=43374736

My wife and I are building Spruce, a Slack app that autocorrects typos in your messages: https://spruce.so/

The motivation behind this was twofold. Firstly, we're both the type of people who often edit our Slack messages (or any other messages like WhatsApp) to correct mistakes. Secondly, my wife is a PMM at a tech company, but she can't code. I've been telling her that anyone can create a decent product with the help of LLMs and tools like Replit, but she usually just rolls her eyes. So, on a flight from London to SF, I suggested we build Spruce. The idea was that she would take the lead and rely on LLMs (ChatGPT + Ollama for when the wonderful Delta internet wasn't working), and she could also treat me as a [more intelligent] LLM. After ~12 hours and a few small arguments, Spruce was born. :)

Some random learnings, top of mind:

1. A non-technical person, even with current LLMs (GPT 4.5), can't go from idea to shipped product without at least some help from a technical person.

2. I now make even more typos because my habit is to type as fast as possible, often without even pressing the space bar. Spruce autocorrects everything in Slack, but when I'm using other products, I feel incomplete.

3. Although I'm practically a native English speaker, it's still my second language. I've added a feature to Spruce, just for myself for now, that DMs me if I make really obvious grammar/English mistakes in my messages. I decided to do this after one of my colleagues, who is very English, said "for prosperity" instead of "for posterity" in several messages, and another said "for all intensive purposes" instead of "for all intents and purposes." It's been a very cool development, at least for me! It's fun because a) it has a nice tone, and b) it's not verbose and only nudges you on major mistakes, and only up to twice a day. Anyway, feel free to give it a try: https://spruce.so/

P.S. This Slack app necessarily needs the permissions it has because it needs to be able to read and edit your messages everywhere (e.g., public channels, private channels, DMs, etc.). Don't install it if that freaks you out. However, it's easy to remove Slack apps, so there's no harm in giving it a try.

Awesome to see this get picked up here! I'm Shahriar, co-founder of Metaview. Any questions, let me know!

There is not :)

Co-founder here. We spend a fair bit of time thinking about this actually. I think the key is to always have something moving proportionally to scroll, otherwise you lose your sense of speed+position and get lost.

Co-founder here. Questions are automatically detected and tidied up. So, we un-waffle. Another consequence of this is that if there's a question like "tell me more about that", we often detect what "that" is referring to and 'expand' the question to something like "tell me more about your last role".

The percentage of interviews with mention of keywords related to the macroeconomic climate (e.g, recession, economy, downturn, uncertainty, inflation) have increased 2x, going from 7% at the start of 2021 to 14% in July 2022.

Here's a shallow but interesting look into this.

I wonder how Svelte would do. Does anyone happen to know?

For sure. There are of course other ways to meaningfully contribute to this boost in productivity as well (e.g., strong async comms) but eventually at some point you need to make a _good_ decision _fast_ — you need context for that.

I think it's certainly highlighted the importance of Consistency. Especially for early-stage teams. Most larger world-class organizations end up having mechanisms, processes, or conventions to ensure context travels [somewhat] fluidly across teams, but smaller companies used to "get away" with not thinking about it as much when everyone was in the same office (read: room). There's now a forcing mechanism which means these things need to be thought of sooner.

Metaview started pre-pandemic and was originally non-remote company, so, we were always deliberate about this even when we didn't "need" to. Having now become a remote team, the transition was coincidentally incredibly easy for us because of our PUP expectations.

I know there are quite a few companies that have strong opinions around this as well. For those who work there, what do you think?

This is interesting. It’s easy to roll your eyes when you see yet another tasks/todo, but the data model innovation here (as it pertains to it integrating across the rest of the stack) is pretty cool — thought-provoking at the very least.

Will give it a go to test the water.

Twitter follow + DM me with mobile number @s16h_ for invite.

Definitely agree with your point about bad management. Do you tend to ask those questions openly or try and figure it out throughout the process?

Check out https://highlight.app/

I've put a bunch of the videos mentioned here into a Highlight list: https://highlight.app/stajbakhsh/doug-engelbart

Feel free to add more to it.

I've moved the content of http://pymust.watch/ to Highlight in an effort to gradually build a better ecosystem around it. Keen to hear thoughts.