yunyu

35 ms·

31.
by yunyu 1y ago
Never understood this take. Just because it doesn’t address the systemic inequities in society doesn’t mean it’s not useful.
32.
by yunyu 1y ago
You're absolutely right! This isn't just X, it's Y...
33.
by yunyu 1y ago
This is just advertorial marketing slop?
34.
by yunyu 1y ago
We solved this at Ramp on the expenses/AP side with an agentic RAG implementation and a custom embedding model, backed by D&B/Google/user-submitted corrections. If curious, details here: https://engineering.ram
35.
by yunyu 1y ago
To be clear, this is what we ended up doing
36.
by yunyu 1y ago
Hey all, member of the benchmark team here! The goal for this project was to see how LLMs well could do bookkeeping without an overly opinionated scaffold. We gave them access to processed transaction records and code execution tools, but i
37.
by yunyu 1y ago
"Your fancy AI scaffolds will be washed away by scale." - Noam Brown
38.
by yunyu 1y ago
A certain level of errors is tolerable/inevitable. But the accountants need to be able to correct for them once they build up
39.
by yunyu 1y ago
No comment, the good news is that accounting and taxes are verifiable - so in principle it is possible to RL models to do them correctly
40.
by yunyu 1y ago
This was true up until they started training them using Reinforcement Learning from Verifier Feedback (started with O1). By sticking a calculator in the training loop, they seem to have gotten out of the arithmetic error regime. That said,
41.
by yunyu 1y ago
LLMs are on the verge of replacing data scientists and investment bankers. But can they perform simple accounting tasks for a real business? We built AccountingBench, a test where LLMs must "close the books" for a real SaaS busine
42.
Can LLMs Do Accounting?(accounting.penrose.com)
6 points by yunyu 1y ago|8 comments
43.
by yunyu 1y ago
Maybe put the warning below the UI, so it doesn't cause the layout to change?
44.
by yunyu 1y ago
If you are a gamer, you are no longer NVIDIA's most important customer.
45.
by yunyu 2y ago
4o mini is a weak, old model. I’m pretty sure even regular o1 (non-pro) would be able to crack a simple vuln
46.
by yunyu 2y ago
On the flip side, this is extremely fast and works pretty well on my M1 Pro Macbook running Chrome. The only downside is that it's a little blurry (as it's rendering at 1x pixel scale).
47.
Show HN: Physically accurate black hole simulation using your iPhone camera(apps.apple.com)
316 points by yunyu 2y ago|115 comments
48.
by yunyu 2y ago
Score of 6. Started and sold a YC startup
49.
by yunyu 3y ago
Congrats on launching – we just spent quite a bit of time replicating/transforming our primary database into clickhouse for OLAP use cases, and it would have been way easier if there were a postgres-native solution. Hoping the managed
50.
by yunyu 3y ago
The above is a common anthropocentric take that has been repeatedly disproven by the last decade of deep learning research: http://www.incompleteideas.net/IncIdeas/BitterLesson.html Understanding audio with inputs in
51.
by yunyu 3y ago
I agree with your point at the highest (pretrained model architect) level, but tokenization/encoding things into the frequency domain are decisions that typically aren’t made (or thought of) by the model consumer. They’re also not stri
52.
by yunyu 3y ago
Vector databases are oriented around search and retrieval. The usual approach of generating vectors is to fine tune a large pretrained model and extract the inner representations. Because the dataset contains successful queries and retrieva
53.
by yunyu 3y ago
I got SMILE instead of LASIK for that exact reason (avoiding dry eye), and have perfect vision with moist eyes. Highly recommend avoiding LASIK if you qualify for newer procedures.
54.
by yunyu 3y ago
Ryan literally just made the previous CEO leave?
55.
by yunyu 3y ago
Jsonformer - https://github.com/1rgs/jsonformer-claude , https://github.com/1rgs/jsonformer
56.
by yunyu 3y ago
Extracting data from raw unstructured text/images and converting them into formatted fields
57.
by yunyu 3y ago
What's the plan/timeline for offering cosine similarity support, given that most OSS embedding models are fine tuned on a contrastive cosine distance objective?
58.
by yunyu 3y ago
You simply sample tokens starting with the allowed characters and truncate if needed. It’s pretty efficient, there’s an implementation here: https://github.com/1rgs/jsonformer
59.
by yunyu 3y ago
Most legitimate CS programs like Stanford or CMU CS are not ABET accredited.
60.
Jsonformer: Generate structured output from LLMs(github.com)
340 points by yunyu 3y ago|83 comments

More ›