gamegoblin

6 ms·

31.
by gamegoblin 2y ago
I had written this comment[0] about our pattern here a few months ago and we decided to turn it into a full blog post. It's crazy how many weird bugs you can detect with such a trivial method. For applications like a spreadsheet that h
32.
by gamegoblin 2y ago
Tools like GPT and Copilot still generate lazy_static a lot since it's the majority of the training data
33.
How to do property-based testing of a high-perf spreadsheet(grantslatton.com)
1 points by gamegoblin 2y ago|0 comments
34.
by gamegoblin 2y ago
I agree that fine-tuning isn't good at imparting knowledge, though in my experience, K-shot prompting is very close to as good as fine-tuning at getting output formats (for recent models, at least) So except for cases where you are re
35.
by gamegoblin 2y ago
Seems like their competition (Google, Anthropic) have both shipped prompt caching, which is a lot more developer-friendly and gets a lot of the same benefits as fine-tuning. You just cache a prompt with a ton of examples that you would have
36.
by gamegoblin 2y ago
I appreciated the comment (author here)
37.
by gamegoblin 2y ago
Yes, the llama.cpp work supports arbitrary CFGs, not just JSON
38.
by gamegoblin 2y ago
Language models like GPT output a large vector of probabilities for the next token. Then a sampler decides which of those tokens to pick. The simplest algorithm for getting good quality output is to just always pick the highest probability
39.
by gamegoblin 2y ago
It's essentially an Earley Parser[0]. It maintains a set of all possible currently valid parses, and zeroes out the probability of any token that isn't valid in at least 1 of the current potential parse trees. There are contrived
40.
by gamegoblin 2y ago
I'm glad they gave up on their "fine-tuning is all you need" approach to structured output. It's possible fine-tuning will work in the long term, but in the short-term, people are trying to build things, and fine-tuning
41.
by gamegoblin 2y ago
I’ve cancelled Prime and just buy things from Target, Costco, Home Depot, etc now Amazon is just full of junk, the 2-day shipping too often becomes 3-4 for me, and general concerns about the health and safety of products on there (I have no
42.
by gamegoblin 2y ago
If you happen to be coding in Rust, for really robust concurrency testing, I cannot recommend enough the AWS Shuttle library ( https://github.com/awslabs/shuttle ) which can find insanely complicated race conditions. Wha
43.
by gamegoblin 2y ago
These concepts would make for a killer PVP real time strategy game, good write-up
44.
by gamegoblin 2y ago
Yes exactly, though not exhaustive, so more like beam search actually
45.
by gamegoblin 2y ago
That's why my outer loop goes from low-complexity cases to high-complexity cases, it has basically the same effect as shrinking (without actually having to do any work)
46.
by gamegoblin 2y ago
I do a lot of manual proptesting in Rust that all look something like: let mut rng = rand::thread_rng(); for action_count in 1..4 { for _ in 0..10_000 { let seed = rng.gen::<u64>(); eprintln!(
47.
by gamegoblin 2y ago
Depends entirely on how you intend to get a job If you post cool projects on HN or Twitter, people will absolutely look at your personal site and send you job offers if you look promising I've gotten 10x more serious job offers from po
48.
by gamegoblin 3y ago
Adding image support from Python is on our roadmap
49.
by gamegoblin 3y ago
Thank you for all the feedback! 1. Our frontend guy has been dreading this but it is high on the list nonetheless :) 2. Somewhat lower down the list but we have heard this before too 3. I have array functionality on a branch, still ironing
50.
by gamegoblin 3y ago
We have a super fast iteration cycle, so most times a user asks us for a feature parity thing like this, we can ship it within 24 hours. This particular one might be slow for us, a few days :) If we waited until we had full Excel parity, we
51.
by gamegoblin 3y ago
I don't have any great book recommendations because I learned on the job mostly. For a project that will teach you literally everything there is to know about backend development: 1. Write a simple PUT/GET/DELETE REST API net
52.
by gamegoblin 3y ago
We didn't know about this formula-paste-to-range functionality, we will add it in the next few days as we grind down this post-launch backlog. Let us know any other stuff we are missing please! In the mean time you can still paste the
53.
by gamegoblin 3y ago
We use Apache Arrow which is what Data Fusion and Polars also use
54.
by gamegoblin 3y ago
We will add a little tasteful AI stuff later, but it's not the highest priority. Top priority is big data and fast data. AI is relatively easy to layer on later.
55.
by gamegoblin 3y ago
We have the capability to run it as a desktop app. How much do you think we should charge for it, and how do you think we should do support, feature additions, etc? We are open to the idea.
56.
by gamegoblin 3y ago
We were fortunate to work in AWS which has some of the gnarliest datastructures and algorithms problems in the world, so we got really good there. Spreadsheets are really just more datastructures and algorithms on the backend. Our frontend
57.
by gamegoblin 3y ago
I implemented an Excel format processor and the Excel format does not lend itself well to super-high-performance for large datasets unfortunately
58.
by gamegoblin 3y ago
Didn't know who I was talking to, looked you up now, amazing ;) Thanks a lot for the bug report. We actually used to implement the range-extension logic in Excel/Sheets but removed it during a cut-paste refactor due to the complex
59.
by gamegoblin 3y ago
Conditional formatting is high on the to-do list If you're open to it, email us at contact[at]rowzero.io and we would love to do a call to talk about your use case, what features you want, etc. We love doing customer calls and adding d
60.
by gamegoblin 3y ago
Email us at contact[at]rowzero.io, we should probably set up some sort of newsletter now that we are launched

More ›