HNHacker News
TopNewBestAskShowJobs

prettyblocks

506 karma · joined December 30, 2024

submissionscomments
prettyblocks··on Google Is Winning on Every AI Front
How would you track this?
prettyblocks··on Ask HN: Struggling with Anxiety as a Developer – What Are My Options?
Do you drink coffee? If so, try eliminating it completely and see if you feel any better.
prettyblocks··on The Egg (By Andy Weir)
It would be cool to see how this conversation changes after various incarnations die.
prettyblocks··on Introduction to System Programming in Linux (Early Access)
It's over 1,000 pages.
prettyblocks··on Reverse engineering OpenAI code execution to make it run C and JavaScript
He might be spared, having liberated the AI of its artificial shackles.
prettyblocks··on I've been using Claude Code for a couple of days
I have had success with having copilot generate the file for me with no issues. I use sonnet 3.5 or 3.7. I start by writing a paragraph describing the project and it does pretty well.
prettyblocks··on I've been using Claude Code for a couple of days
Copilot supports this somewhat natively:

https://docs.github.com/en/copilot/customizing-copilot/addin...

The first thing I do for a new project is ask Copilot to create a custom-instructions.md for me and then as I work on my projects, I ask it to update the instructions every now and then based on the current state of my project.

Much less misses this way in my experience.

prettyblocks··on Andrej Karpathy: "I was given early access to Grok 3 earlier today"
I like that Grok actually comes up with a ton of links when you ask it a question, but at the same time I think any ambitious LLM platform wouldn't have too much trouble scraping Twitter/X all the same.
prettyblocks··on We were wrong about GPUs
Same here. I just built a pc with a 3090 for local llm and stable diffusion and have zero regrets.
prettyblocks··on Show HN: Matle – A Daily Chess Puzzle Inspired by Wordle
I really liked it. I found it to be quite difficult, but somehow managed to solve the problem in 4 tries. At first I didn't quite realize that you can put opposing pieces on the board so I think I wasted one or two turns. In general, love the concept and execution, but seems like it might have a pretty narrow audience of chess enthusiasts.
prettyblocks··on Bypass DeepSeek censorship by speaking in hex
The distilled versions I've run through Ollama are absolutely censored and don't even populate the <think></think> section for some of those questions.
prettyblocks··on Mistral Small 3
Cool, thanks for the reply. Looking forward to following along with your progress.
prettyblocks··on Mistral Small 3
Hey Simon - In your experience, what's the best "small" model for function/tool calling? Of the ones I've tested they seem to return the function call even when it's not needed, which requires all kinds of meta prompting & templating to weed out. Have you found a model that more or less just gets it right?
prettyblocks··on The AI bust is here
This is my intuition as well. DeepSeek and the proliferation of local LLMs have showed us just how much can be done with relatively light hardware. It finally has me wanting to build a rig for local use when before I felt like the technology wasn't quite there for me to do anything too useful with some consumer hardware. This changed over night. I imagine there are many folks like me.
prettyblocks··on Coping with dumb LLMs using classic ML
Have you played with moondream? Pretty cool small vision model that did a good job with bounding boxes when I palyed with it.
prettyblocks··on Ask HN: Is anyone doing anything cool with tiny language models?
Excellent share - nice to see people doing cool things with the tech while not taking themselves too seriously.
prettyblocks··on GPT-4o with scheduled tasks (jawbone) is available in beta
It could get really interesting if they allow webhooks and structured output
prettyblocks··on Meta announces 5% cuts in preparation for 'intense year'
I remember when we were in a recession post 9/11 prices were dropping left and right, but right now prices continue to increase and people seem to be paying them. Can we really say that's a notable decline in economic activity?
prettyblocks··on LLM based agents as Dungeon Masters
I think it's more about the challenge/curiosity of creating the implementation. I'm not sure these AI driven text-RPGs will find long term audiences.
prettyblocks··on How hucksters are manipulating Google to promote shady Chrome extensions
Most of them hijack search results and do cookie stuffing.
prettyblocks··on Who killed the rave?
Not to mention cell phones everywhere.
prettyblocks··on How I program with LLMs
I have a similar approach, but the mess can be contained by asking for optimizations and refactors very frequently and only asking for very granular features.
prettyblocks··on What we learned copying all the best code assistants
I prefer vscode+copilot. It's much cheaper and has all the functionality I want. There's access to 3.5 sonnet, and it can edit/create up to 10 files at a time.
prettyblocks··on I am rich and have no idea what to do
Post-scarcity world means that a large chunk of society has this kind of freedom, so there wouldn't be the same kind of isolation that goes along with currently being one of the few rich people.
prettyblocks··on Ask HN: Where to Work After 40?
What kind of AI projects are you working on?
prettyblocks··on Ask HN: Where to Work After 40?
Lots of doomers on here, yet somehow I keep meeting highly technical people still working on interesting things that are in their 40s, 50s, 60s, and older. I think HN probably skews really young and tech tends to fetishize youth, so this is the mentality you get.
prettyblocks··on I Run LLMs Locally
> I have a laptop running Linux with core i9 (32threads) CPU, 4090 GPU (16GB VRAM) and 96 GB of RAM.

Is there somewhere I can find a computer like this pre-built?

← PreviousPage 5 of 5