afro88

8 ms·

121.
by afro88 10mo ago
It's also extremely good at describing what code is doing, architecture, extrapolating why something is done a certain way etc. Invaluable for me for learning how unfamiliar code works in unfamiliar languages
122.
by afro88 10mo ago
This is exactly the right approach IMO. You find the signal amongst the slop, and all your colleagues see is a better PR.
123.
by afro88 10mo ago
I was a heavy user, but stopped using it mid 2024. It was essentially providing codebase context and editing and writing code as you instructed - a decent step up from copy/paste to ChatGPT but not working in an agentic loop. There was
124.
by afro88 10mo ago
> The key insight from this benchmark is using "human-equivalent hours" rather than actual AI execution time. It's measuring capability complexity, not speed. > What's interesting is the 50% vs 80% reliability gap.
125.
by afro88 10mo ago
Aider wasn't really an agentic loop before Claude Code came along
126.
by afro88 10mo ago
A lot of tools these days put an extra <system> message into the conversation periodically that the user never sees. It fights against context rot and keeps important things fresh.
127.
by afro88 10mo ago
You're not missing much. You can generally use Cursor like Claude Code for normal day to day use. I prefer Cursor because I like reviewing changes in an IDE, and I like being able to switch to the current SOTA model. Though for more au
128.
by afro88 10mo ago
> That's already a huge breach in comparison to mullvad privacy page. And the "3 data points, that's it" of the blog post
129.
by afro88 10mo ago
People say this kind of thing a lot, but in reality the concept of "software engineer" will change and there will still be experience levels with different expectations
130.
by afro88 10mo ago
> Literally everything else the LLM generated on demand for me based on its existing training data, offline That's a ton of scope for hallucinations, surely?
131.
by afro88 10mo ago
Great idea and execution. But please don't get an LLM to write the whole blog post for you. It took the enjoyment out of reading it for me. I want to read about your project in your words. Reading it in LLMs words feels disconnected an
132.
by afro88 10mo ago
> I usually have to leave the happen or the session terminates Assuming you meant "leave the app open", I have the same frustration. One of the nice things about the ChatGPT app is you can fire off a req and do something else.
133.
by afro88 11mo ago
Do you really think that is what the author meant? They're not saying "you only need 1". They're talking about going from 0 to 1. It's sometimes a long and arduous 0. Your framing doesn't help someone trying to
134.
by afro88 11mo ago
I don't need to install an app to then purchase a subscription on pretty much any web app on the internet. You want to make conversion as frictionless as possible. Making potential users install another app to make a purchase will lose
135.
by afro88 11mo ago
I think you'll be waiting a while for the "crashing down". I was a kid when manufacturing went off shore and mass production went into overdrive. I remember my parents complaining about how low quality a lot of mass produced
136.
by afro88 11mo ago
Often the cause is things that most people can handle, without being able to easily wield the tools to handle them. Having a pill that dulls the symptoms gives space to learn and become adept at the tools
137.
by afro88 11mo ago
Yep. A lot of things looked magical in the GPT-4 days. Eventually you realised it did it by chance and more often than not gets it wrong
138.
by afro88 11mo ago
Do we need this, when we have tools like Claude Code, Codex etc that you can talk to about the codebase they are started in?
139.
by afro88 1y ago
I was always disappointed by the Cursor version because the agents would make entirely new mistakes that Cursor IDE wouldn't make locally. Like so much that it was totally unusable. Completely messing up code edits to the point where a
140.
by afro88 1y ago
This actually reads like it was written by one of those character LLMs
141.
by afro88 1y ago
Nothing certain, but I reckon it'll be a while: https://kotlinlang.org/docs/components-stability.html#stabil... The other big problem is debugging. It's impossible to breakpoint kotlin when debugging from swi
142.
by afro88 1y ago
> Self-hosted video isn't that hard but no one makes a baby-monitor centric solution It sounds like they're not hosting it though. They are processing it, and storing it temporarily while it's queued. A fully self hosted A
143.
by afro88 1y ago
24/7?
144.
by afro88 1y ago
We use it in my team and it works well enough, but iOS is a bit second class citizen. Everything translates to Obj-C (NSObject at the root), so even something as simple as a data class becomes NSObjects with a cumbersome dev experience rath
145.
by afro88 1y ago
Think of this more like Kotlin Multi Platform. Shared logic, not UI.
146.
by afro88 1y ago
Doesn't staging need to be a (downsized) replica of prod, infra wise to give confidence that changes will be stable and working in prod? Genuine question.
147.
by afro88 1y ago
> Isn't this just an emulator or rebadged something-or-other? >The Commodore 64 Ultimate from the only original Commodore® brand (est. 1958) is brand new hardware-based Commodore 64 technology. It features SID chip-reactive LEDs
148.
by afro88 1y ago
> If your company is in anything finance-adjacent or critical infrastructure GP said: > most companies Most companies aren't finance-adjacent or critical infrastructure
149.
by afro88 1y ago
All the things you listed are irrelevant for a touring musician. All they care about is that equipment is used that sounds good. If it's a pain to set up or manage, that's the roadie and live mix engineer's job. Not saying so
150.
by afro88 1y ago
It's quite literally the opposite. Simon is tracking how well the "pig" dances as each model gets better (or worse) at it

More ›