jermaustin1

5 ms·

1.
by jermaustin1 5d ago
Not in my experience. Multiple submissions (non Show HN) of mine have been second chanced. I can't be certain, but I don't think any of my Shows have been... I guess that says more about the content I create than the content I sha
2.
by jermaustin1 5d ago
They already do this. It's called the "second chance" pile. You will get an email saying that your submission didn't get the love they felt it should, and will add it to a list of submissions that get submitted to the fr
3.
by jermaustin1 11d ago
A bespoke design like that a year ago would have cost $200-2000+ between design and development. Hell, it probably still does unless the person wanting the design is already a developer who knows how to prompt. Everything costs more now tha
4.
by jermaustin1 13d ago
I worked for a different Steve Ross (Dolphins owner, Related Companies, etc), and we did this in house. From 2008-2017, “we” (I was a contractor 08-09, and employee 15-17) basically turned every spreadsheet into a web application internally
5.
by jermaustin1 13d ago
Years ago, I lived in NYC, and my roommate was a director of photography for National Geographic, and various other nature documentaries. I loved photography (still do, but much less time for it as a late 30s adult than a mid 20s adult), an
6.
by jermaustin1 18d ago
Opposite policy at one of my clients (kind of). I am responsible for the code that upper management's Claude produces. Some Mondays, I will start work with a half dozen emails with attachments of Claude generated code for something I d
7.
by jermaustin1 18d ago
Wow! I'm utterly speechless. I love space "browsers" - I lose myself in them, just clicking through all the entities. Also TIL that there are satellite galaxies, and I've never noticed those on other toys/tools like
8.
by jermaustin1 19d ago
I've done pretty decent local prose->json extraction using Qwen and Phi and Gemma. I'm sure most of it comes down to prompts, and all of them run over 100tps on a 3090. Smaller cards will likely be slower, but Qwen3.5 9B is sma
9.
by jermaustin1 20d ago
They cost more to run than hosted anyway. But that isn't the point of having them. They are a playground, a backup when the internet is down, or claude is down. They can render Blender scenes pretty well. They play any game I want. You
10.
by jermaustin1 20d ago
I don't think there is tunnel vision. I'm just saying that I have a couple 3090s I invested in a handful of years ago, and they are still going strong today as multiple GPU-needing technologies emerged. I'm not saying everyon
11.
by jermaustin1 20d ago
Having multiple 6 year old cards doesn't seem like it's that big of burden for local LLMs. I get that a lot of people don't have them. And a single one can be VERY performant. And the smaller models like a 7B can run on much
12.
by jermaustin1 20d ago
To me, most local models work just fine for anything you can be patient for. If I want something quicker, I will go to a SOTA model via API, but with multiple 3090s, I have never really needed a hosted model for a lot of my experiments. For
13.
by jermaustin1 20d ago
I know this is mostly true at huge companies, but having spent my career at various "small" corporations, with maybe a few thousand employees total (and only ~50 of which were in IT). I was always surprised when the CEO knew names
14.
by jermaustin1 22d ago
I just wait for a cycle or two where the leaps and bounds are more like hops and steps. So if the M7 Ultra improves inference by 2x over the M5, but the M9 Ultra only improves by 1.2x over the M7, that's my signal to buy. Unfortunately
15.
by jermaustin1 22d ago
I've been holding out, because I think my next purchase will be a Studio with an Ultra Chip in it. I'm wanting it to be a "forever" server, so I'm holding out while I can.
16.
by jermaustin1 23d ago
Have you actually read your README? There are so many parts that are just "words" thrown together to sound technical. It reminds me of books you pick up in a science fiction game.
17.
by jermaustin1 23d ago
What did I just read? I'm 99.9999% certain that 75+% of that README was LLM hallucination.
18.
by jermaustin1 23d ago
I did a similar thing for the UK when prepping to drive around. I hooked my steering wheel up to Forza Horizon 4, and just drove slow cars on the wrong side of the road through Edinburgh, following as many of the road laws as I knew. It was
19.
Show HN: HN Notify – Subscribe to HN replies, users, keywords, and domains(hnnotify.org)
3 points by jermaustin1 23d ago|0 comments
20.
by jermaustin1 24d ago
I was the first of my CURRENT work colleagues to use it, but I’m in a lower tech field.
21.
by jermaustin1 26d ago
I've often wondered if US Debt has ever actually mattered for the US. They were complaining when Bush took Clinton's surplus and wasted it on wars, and grew the deficit to $10T (some was Great Recession bailout related). The compl
22.
by jermaustin1 27d ago
Here's the best way I can explain how I "think" because it is mostly passive, but sometimes, I am actively thinking. If I'm asked a hard question, like what city I'd most like to visit, I kind of have to look up cit
23.
by jermaustin1 27d ago
How good does it filter out the "out of context" audio? My party's sessions are 4-6 hours once a month, an I would say only around 45-90 minutes are the actual campaign, but spread across that entire time.
24.
by jermaustin1 27d ago
Because my brain thinks through text in a forward motion. Pausing at the end of each word and searching for the next. My entire brain runs on sentences and words since I have no inner eye or whatever. So my thinking and writing both work ki
25.
by jermaustin1 27d ago
I'm very interested in Diffusion text models. The concept of taking noise and adding words starting randomly all over the response, and filling in the noise from there on breaks my brain. I'm sure I have a fundamental misunderstan
26.
Show HN: HN Notify(hnnotify.org)
3 points by jermaustin1 28d ago|0 comments
27.
Dear researchers: Is AI all you've got?(austinhenley.com)
7 points by jermaustin1 28d ago|0 comments
28.
by jermaustin1 29d ago
Love this so much. I always get in my own way with making a game like this. I keep trying to add too much instead of it just being an incremental clicker game.
29.
by jermaustin1 1mo ago
This just reminded me to cancel it. I got pretty annoyed back in March when I did a comparison between Cursor and Claude Code to see if Claude had gotten better, and my cursor burned $200 unsupervised on extra credits in a single session. I
30.
by jermaustin1 1mo ago
That is interesting. I'll have to test that theory out today.

More ›