theOGognf

25 ms·

1.
by theOGognf 10mo ago
I feel like both this comment and the parent comment highlight how RL has been going through a cycle of misunderstanding recently from another one of its popularity booms due to being used to train LLMs
2.
by theOGognf 1y ago
I really enjoyed it. Very Hitchhiker’s Guide -like. It’s been a while since I’ve read something like that. Thanks for sharing
3.
by theOGognf 1y ago
I used to follow advancements in RL pretty closely since 2016; it’s cool to see how far methodologies and algorithms have come to be able to do complex tasks completely offline. Dreamer 4 is another big leap in the Dreamer series
4.
Dreamer 4(arxiv.org)
3 points by theOGognf 1y ago|1 comments
5.
by theOGognf 1y ago
I didn’t see this before, but I work at a non-profit, government hospital system. So increasing profits, although probably a good answer most of the time, is probably not as applicable in this case lol
6.
by theOGognf 1y ago
Along the same lines of missing good junior engineers at work, we occasionally interview stellar engineers that’ve inflated their resume a bit to get an interview, but we end up rejecting them for not having all the specific experiences our
7.
by theOGognf 1y ago
Not sure about other hospital systems, but the one I work at is developing CV systems to help fill workforce gaps in places where there isn’t as many trained professionals or even resources to train professionals
8.
by theOGognf 1y ago
No, that is the norm. Radiologists speak with their colleagues the most, and patients rarely
9.
by theOGognf 1y ago
This article is pretty good. My current work is transitioning CV models in a large, local hospital system to a more unified deployment system, and much of the content aligns with conversations we have with providers, operations, etc.. I thi
10.
by theOGognf 1y ago
It’s ironic that HN threads, arguably one of the forums where the majority of users should understand a tech company and its tech, about Palantir always devolve into some weird speculative and conspiracy-like discussion. Palantir’s docs are
11.
by theOGognf 1y ago
I’ve used Diesel for a bit now but haven’t had issues wrangling the type system. Can you give an example of an issue you’ve encountered?
12.
by theOGognf 1y ago
Is it common to see Metaflow used alongside MLflow if a team wants to track experiment data?
13.
by theOGognf 1y ago
Along with burning tokens, how MCP servers are ran and managed is resource wasteful. Running a whole Docker container just to have some model call a single API? Want to call a small CLI utility, people say to run another Docker container fo
14.
by theOGognf 1y ago
That is true too
15.
by theOGognf 1y ago
Right, this is only power usage. Factoring in labor and all that would make it more expensive for sure. However, it’s not like it’s a complex system to maintain. We use a popular inference server and just run it with some modest rate limits
16.
by theOGognf 1y ago
Some anecdotal data, but we recently estimated the cost of running a LLM at $WORK by looking at power usage over a bursty period of requests from our internal users and it was on the order of $10s/mil tokens. And we arent a big place,
17.
by theOGognf 1y ago
I was wondering this myself and made a tool-calling endpoint with OpenAPI to see if it worked well. Turns out it does: https://github.com/theOGognf/toi
18.
Show HN: An extensible personal assistant server(github.com)
5 points by theOGognf 1y ago|0 comments
19.
by theOGognf 1y ago
It’s funny seeing this blog post again. This is actually a reference I used to make a poker game as a state machine last year: https://github.com/theOGognf/private_poker It made the development feel a lot safer and it’
20.
by theOGognf 1y ago
This looks like an article about the recent Nature publication. Was confused at first because DreamerV3 is a couple of years old now
21.
by theOGognf 2y ago
Hey Justin. First time coming across one of these postings. I saw a role that matched up to what I do now and applied (Sr MLE).
22.
by theOGognf 2y ago
By hitting it sufficiently hard
23.
by theOGognf 2y ago
They do this in subsets of aerodynamics for optimizing the tradeoff between simulation speed and fidelity in cases where there will be a lot of runs with the same geometry but different parameters
24.
by theOGognf 2y ago
Ive come to like coding in Rust a lot more than any other language Ive tried. It is really nice having a ton of tooling out of the box that helps the development experience (formatting, linting, rich compiler errors and suggestions, etc). I
25.
Show HN: Poker over SSH(github.com)
10 points by theOGognf 2y ago|3 comments
26.
by theOGognf 2y ago
Ironically, that sentence is pretty plain for a dynamics and control paper lol
27.
Show HN: Finagg – free and nearly unlimited financial data(github.com)
11 points by theOGognf 3y ago|0 comments
28.
Show HN: An end-to-end reinforcement learning library for infinite horizon tasks(github.com)
1 points by theOGognf 3y ago|0 comments