ZainRiz

I'd respectfully push back on the framing here.

If you look at value as purely the LLM output, then there's a valid argument that the best frontier models will always be better than fine tuned specialists. (I'm not convinced personally, but it's a defensible claim)

But that misses two dimensions: 1. The cost of acquiring that output 2. What is actually "good enough" for that specialist domain

Not every output needs to be the best to produce value.

And as specialist models increase in cost, their cost/value proposition goes down.

At some point, there's a threshold where cheaper, fine tuned models are "good enough" at the task and also substantially cheaper than the expert models.

That's where fine tuning helps.

Personally, I became a believer in fine tuning after fine tuning a 1B Qwen model as a second pass over my local voice transcription app, achieving excellent accuracy at ~zero token cost and waaaay lower latency than if I'd invoked my Claude subscription under the hood.

If you're in finance, you earn enough to live in NYC

> imagine what happens to NYC when all the people doing the menial work keeping the city alive (have to) move away? Whoops, now everyone is going to drown in trash and feces!

That's not a "the poor middle class folks in NYC need help" story, that's a "the rich folks of NYC need folks to serve them" story. They're welcome to strategize however they like to incentivize people working there. Manipulative heartstring tug are not welcome however.

Maybe now someone will finally fix the bug that causes claude code to randomly scroll up all the way to the top!

They also have a "keep going" keyword, literally just "continue" or "keep going", just for logging.

I've been using "resume" this whole time

This is so cool! Now, I didn't notice anything about waterproofing on your website.

Any plans to come out with a version I don't have to worry about getting wet?

Avoid casing entirely and use single letter function and variable names

Case insensitive, of course

Same here. I found the biggest beneficiary of a design doc was my own thinking process (Xoogler here)

Google has the concept of a mini-design doc called a "One-pager"

In 2018, the template for this "One-pager" was two pages long

With Microsoft they tell you years in advance when the product will reach EOL, so you have plenty of time to prepare

On top of that, most of their products are software running on your desktop and will still keep working after EOL

And if you're a government or enterprise with spectacularly big wallets, you can choose to pay Microsoft big $$ to keep shipping you additional security patches

I think the claim is that some other company in the robot vacuum business can buy iRobot instead of Amazon and get that sweet sweet IP

+1. At 22, working as a full time software engineer at a tech company, I still didn't think I knew how to do anything useful

Even though I've drawn pictures, built toys and contraptions contraptions out of wood, made a rube golderg machine

Still...

The thing I'm the proudest about...

Is that time I opened a jammed coaxial cable with nothing but two butter knives.

The challenge with books like The God Delusion is that it's barely one step above strawman arguments.

It attacks claims made by some simple lay people, based on their half-baked understanding of their own faiths (which they never really dug into). However, people well versed in the religion don't actually make any such claims.

The arguments being debunked are ones that religious scholars themselves would have disagreed with. Such misrepresentations don't guide anyone

Kindly also report the usage of "Chai Tea" ('Chai' means 'Tea' in Pakistan/India, which is where that flavor originated from)