calebkaiser

8 ms·

61.
by calebkaiser 8mo ago
There is a platform called ethical ads for developer focused advertising: https://www.ethicalads.io/
62.
Agent Optimizer: Self-improving prompts from production data(github.com)
1 points by calebkaiser 8mo ago|0 comments
63.
by calebkaiser 8mo ago
I work with/am friends with many junior-ish developers who are in the same place as you (got into programming in their late 20s around the 2020 hiring cycle). I'm very sorry for the stress you're dealing with. I don't kn
64.
Deep Implicit Layers(implicit-layers-tutorial.org)
1 points by calebkaiser 9mo ago|0 comments
65.
by calebkaiser 9mo ago
If anyone is curious, Beads is an agent memory project from the same developer: https://github.com/steveyegge/beads
66.
Show HN: Opik Optimizer – open-source agents for self-improving LLM applications(github.com)
2 points by calebkaiser 9mo ago|0 comments
67.
Opik Agent Optimizer – Open-Source Prompt Optimization Framework(github.com)
6 points by calebkaiser 10mo ago|1 comments
68.
by calebkaiser 11mo ago
Hello friend!
69.
by calebkaiser 11mo ago
I don't really understand this line of criticism, in this context. What would "generalizing" the information in this article mean? I think the author does a good job of contextualizing most of the techniques under the general
70.
by calebkaiser 11mo ago
I think it's fair to question the use of the term "engineering" throughout a lot of the software industry. But to be fair to the author, his focus in the piece is on design patterns that require what we'd commonly call s
71.
by calebkaiser 11mo ago
Most of the inference techniques (what the author calls context engineering design patterns) listed here originally came from the research community, and there are tons of benchmarks measuring their effectiveness, as well as a great deal of
72.
by calebkaiser 11mo ago
Any of the "design patterns" listed in the article will have a ton of popular open source implementations. For structured generation, I think outlines is a particularly cool library, especially if you want to poke around at how co
73.
by calebkaiser 11mo ago
Based on the comments, I expected this to be slop listing a bunch of random prompt snippets from the author's personal collection. I'm honestly a bit confused at the negativity here. The article is incredibly benign and reasonable
74.
by calebkaiser 1y ago
There's been a decent chunk of research in this direction over the years. Michael O'Boyle is pretty active as a researcher in the space, if you're looking for stuff to read: https://www.dcs.ed.ac.uk/home/
75.
by calebkaiser 1y ago
LoRa uses singular value decomposition to get the low rank matrices. In different optimizers, you'll also see eigendecomposition or some approximation used (I think Shampoo does something like this, but it's been a while).
76.
by calebkaiser 1y ago
There's a lot of great work both around supporting memory efficient inference (like on a closer-to-consumer machine), as well as on open source code-focused models. A lot of people are excited about the Qwen3-Coder family of models: h
77.
by calebkaiser 1y ago
I don't think this is an AI specific thing. I work in the field, and so I'm around some of the most enthusiastic adopters of LLMs, and from what I see, engineering cultures surrounding LLM usage typically match the org's prev
78.
by calebkaiser 1y ago
The 1890s were the launching point for widespread unionization among coal miners in places like my home-state of Kentucky. Company towns were increasingly common, and major motivations for unionization were to combat things like being paid
79.
Why Neural Networks Can Discover Symbolic Structures(arxiv.org)
9 points by calebkaiser 1y ago|0 comments
80.
by calebkaiser 1y ago
Neel Nanda is also very active in the field and writes some potentially more approachable articles, if you're interested: https://www.neelnanda.io/mechanistic-interpretability Much of their work is focused on discover
81.
by calebkaiser 1y ago
You might be interested in work around mechanistic interpretability! In particular, if you're interested in how models handle out-of-distribution information and apply in-context learning, research around so-called "circuits"
82.
by calebkaiser 1y ago
Typically, you would collect a ton of execution traces from your production app. Annotating them can mean a lot of different things, but often it means some mixture of automated scoring and manual review. At the earliest stages, you're
83.
by calebkaiser 1y ago
I'm biased in that I work on an open source project in this space, but I would strongly recommend starting with a free/open source platform for debugging/tracing, annotating, and building custom evals. This niche of the field
84.
by calebkaiser 1y ago
I think this is recency bias. There was roughly 33 months between the initial release of GPT 3 and GPT 4, which was released in March 2023.
85.
by calebkaiser 1y ago
Before getting too invested in any conclusions drawn from this piece, it's important to recognize this is mostly PR from Chatterbox. From the Chatterbox site: > Our patented AIMI platform independently validates your AI models &
86.
by calebkaiser 1y ago
Yeah that particular aspect of his claim feels like the weakest. Here is a slightly more in-depth piece where a wildlife biologist mentions other possible forcing functions that cause the snakes to eat the newts: https://baynatur
87.
by calebkaiser 1y ago
I think the preceding sentence in that paragraph answers it. Important context here is that garter snakes tend to swallow prey whole. tldr: a strongly poisoned newt survives consumption attempts. > And it explains why the newts keep evol
88.
by calebkaiser 1y ago
This is awesome! I've recently begun diving deeper into working with grammars, using them as part of a new project, and these tools look super useful.
89.
by calebkaiser 1y ago
Awesome! I appreciate all the open work your team does. A couple years ago, I was staffed on a project that was adopting RSC super early on, and I vividly remember crawling through Shopify blogs/code as one of the few solid resources
90.
by calebkaiser 1y ago
I'm a big fan of web components, and this seems like a very cool project. I'm curious about how it fits into the broader frontend ethos at Shopify. I remember the Shopify team being one of the earliest proponents of React Server C

More ›