bravura

9 ms·

91.
by bravura 2y ago
People pay attention to canaries in coal mines for good reason
92.
by bravura 2y ago
Correlation of 0.25-0.5 being poor is very problem dependent. For example, in difficult perceptual tasks ("can you taste which of these three biscuits is different" [one biscuit is made with slightly less sugar]), a correlation of
93.
by bravura 2y ago
We're using it to RCA infrastructure incidents.
94.
by bravura 2y ago
My comment was about "why would cloudflare make a fluffy blog post like this?" And I believe its to attract people to Lisbon. I never said that I support this oversimplification of Portugal. Which is why I live in a small town, so
95.
by bravura 2y ago
The task: I have a massive OpenAPI spec, and I want to drop all operations except for read-only ones. I want to preserve all referenced types, but slim the YAML as much as possible and remove extraneous elements. I've tried prunes oper
96.
by bravura 2y ago
I'm looking for a solution to filter large openapi specs to the most concise complete subset. I've implemented three different variations, two of which appear not to prune enough, and one of which appears to prune too much. Any re
97.
by bravura 2y ago
Subtext: In case a) you want to move to Europe, b) cannot tolerate weather that differs from California, c) like Golden Gate bridges, are you aware that Cloudflare has a Lisbon office? Source: I live in Portugal and notice all the discussio
98.
by bravura 2y ago
Okay I think I qualify. I'll bite. LeCun's argument is this: 1) You can't learn an accurate world model just from text. 2) Multimodal learning (vision, language, etc) and interaction with the environment is crucial for true l
99.
by bravura 2y ago
For web search, also consider the Kagi and Tavily APIs.
100.
by bravura 2y ago
What a smear. A lot of applied work in vision and audio glues together different existing modules, instead of training the whole thing end-to-end. In an ideal world, things are ideal. But the world isn't a grad student's wet dream
101.
by bravura 2y ago
Yes but Steve Jobs also understood the paradox of choice, and the importance of having incredibly clear delineation between every different product in your line.
102.
OlmOCR: Open-source PDF to plain text, $190 USD / million pages
1 points by bravura 2y ago|0 comments
103.
by bravura 2y ago
Huh, I suspect you completely missed the point of the quote. Jobs was talking about how people don't know what they want until you show them. He was fighting in reaction to the idea that what people say they want will guide you in the
104.
by bravura 2y ago
Is this it? From wikipedia: 'A studio version was made for John Peel's BBC Radio 1 show at the Maida Vale Studios on 28 October 1977 and first broadcast soon after on 14 November. It includes a separate section that begins with an
105.
by bravura 2y ago
I think you missed my point. One of the reasons it's difficult to get good employment with an EU-based company is because it's also harder to terminate you. Germany: Requires extensive justification for firing, works council appro
106.
by bravura 2y ago
You seem to be unaware that worker protections against layoffs can be much stronger in many European countries. [edit: Also, you can always move back to the US and get a new job.]
107.
by bravura 2y ago
If you are working at bigco and want to transfer to EU: * You persuade your manager that this is okay and name the country. * BigCo Inc. terminates you and BigCo Lda. or Srl. or whatever there subsidiary in that country makes you an offer l
108.
by bravura 2y ago
https://aider.chat/2024/09/26/architect.html "Aider now has experimental support for using two models to complete each coding task: An Architect model is asked to describe how to solve the coding problem
109.
by bravura 2y ago
Sorry, not to belabor this point. Would that suggest to you what you did wrong? Or purely show you what you got right? How chatty is this variable?
110.
by bravura 2y ago
True but if can write an LLM in 100 lines of JAX, perhaps we should just train that LLM to parse JSON?
111.
by bravura 2y ago
To the poster who wrote: "Hey Saurabh, will you be willing to teach me this on a call? I'm willing to pay for it (im not rich, so, dont expect much please). I will be having a lot of questions, mostly related to core concepts of t
112.
by bravura 2y ago
Is there any automatic way to get warned against these antipatterns?
113.
by bravura 2y ago
In the grand scheme of things, is the source code that valuable if it's released in a hurry? For a niche product like this, most likely it will be hot pile of garbage.
114.
by bravura 2y ago
Aside, but what do you recommend for internal docs? Particularly if your internal docs are all Markdown?
115.
by bravura 2y ago
Depends. That's true WHEN streaming services just work. But when they don't work, you're SOL with them and they just won't take your damn money. e.g. you live in EU but the account home is the US, or you live in Portugal
116.
by bravura 2y ago
A time-traveling debugger for Python + LLM would be amazing.
117.
by bravura 2y ago
Check out "Attention as an RNN" by Feng et al (2024), with Bengio as a co-author. https://arxiv.org/pdf/2405.13956 Abstract: The advent of Transformers marked a significant breakthrough in sequence modelling,
118.
by bravura 2y ago
I am super interested in sidecar. When will it support o3-mini-high? I also need an SDK to script it, tbh. What I want is to have actually a few different agents that interact with each other. Do you expose a good SDK?
119.
by bravura 2y ago
"[LLM-generated code] is ultra-useful if you can close the loop between the LLM and the execution environment (with an “Agent” setup). There’s lots to say about this, but for the moment: it’s a semi-effective antidote to hallucination:
120.
by bravura 2y ago
I appreciate your argument, but add the following nuance: Latent space reasoning can represent and manipulate UNCERTAINTY more concisely and elegantly than token space reasoning.

More ›