HackerTrans
TopNewTrendsCommentsPastAskShowJobs

rckrd

no profile record

Submissions

Rebuilding Coginition's Agentic MapReduce

mattrickard.com
2 points·by rckrd·12 days ago·0 comments

Rwx: "Ralph Wiggum Loop" util for Claude/codex

github.com
2 points·by rckrd·6 months ago·0 comments

Context-Free Grammar Parsing with LLMs

matt-rickard.com
1 points·by rckrd·2 years ago·0 comments

Every Sufficiently Advanced Configuration Language Is Wrong

matt-rickard.com
2 points·by rckrd·2 years ago·0 comments

The Problems with "Cloud-Prem"

matt-rickard.com
3 points·by rckrd·3 years ago·0 comments

Copilot Is an Incumbent Business Model

matt-rickard.com
3 points·by rckrd·3 years ago·0 comments

The Model Is Not the Product

matt-rickard.com
1 points·by rckrd·3 years ago·0 comments

The Cost of Index Everything

matt-rickard.com
4 points·by rckrd·3 years ago·0 comments

What If Google Wasn't the Default?

matt-rickard.com
26 points·by rckrd·3 years ago·53 comments

The Context Length Observation

matt-rickard.com
1 points·by rckrd·3 years ago·0 comments

Infrastructure as Code Will Be Written by AI

matt-rickard.com
4 points·by rckrd·3 years ago·1 comments

On Mixing Client and Server

matt-rickard.com
2 points·by rckrd·3 years ago·0 comments

When A/B Testing Doesn't Work

matt-rickard.com
1 points·by rckrd·3 years ago·0 comments

The Inner-Platform Effect

matt-rickard.com
1 points·by rckrd·3 years ago·0 comments

Why Is the Front End Stack So Complicated?

matt-rickard.com
73 points·by rckrd·3 years ago·68 comments

A List of Leaked System Prompts

matt-rickard.com
4 points·by rckrd·3 years ago·1 comments

Anticipate the Cheap

matt-rickard.com
1 points·by rckrd·3 years ago·0 comments

Is Data Still a Moat?

matt-rickard.com
1 points·by rckrd·3 years ago·0 comments

Incentives Behind Programming Languages

matt-rickard.com
2 points·by rckrd·3 years ago·0 comments

LLMs as System 1 Thinkers

matt-rickard.com
2 points·by rckrd·3 years ago·0 comments

comments

rckrd
·3 years ago·discuss
Interesting that none of the new features (DALLE-3, Advanced Data Analysis, Browse with Bing) are usable without enabling history (and therefore, using your data for training).
rckrd
·3 years ago·discuss
I've also compiled a list of leaked system prompts from various applications.

[0] https://matt-rickard.com/a-list-of-leaked-system-prompts
rckrd
·3 years ago·discuss
Logit-bias guidance goes a long way -- LLM structure for regex, context-free grammars, categorization, and typed construction. I'm working on a hosted and model-agnostic version of this with thiggle

[0] https://thiggle.com
rckrd
·3 years ago·discuss
We use a similar trick and expose it via an API. Much easier to parse when you can guarantee the shape of the output

[0] https://thiggle.com/
rckrd
·3 years ago·discuss
We've found the same. A lot of usage through our LLM Categorization endpoint. The toughest problem was actually constraining the model to only output valid categories and not hallucinate new ones. And to only return one for single-classification (or multiple if that's the mode).

[0] https://matt-rickard.com/categorization-and-classification-w...
rckrd
·3 years ago·discuss
In more impressive news, "38% of code generated by GPT-4 does not contain API misuses"
rckrd
·3 years ago·discuss
I also released a hosted version of my open-source libraries ReLLM and ParserLLM that already supports APIs for

* Regex completion for LLMs

* Context-free Grammar completion for LLMs

https://thiggle.com/

[0] https://github.com/r2d4/rellm

[1] https://github.com/r2d4/parserllm

[2] https://github.com/thiggle/api

There's also another API on Thiggle that I've build that supports classification via a similar logit-based strategy.
rckrd
·3 years ago·discuss
I just released a zero-shot classification API built on LLMs https://github.com/thiggle/api. It always returns structured JSON and only the relevant categories/classes out of the ones you provide.

LLMs are excellent reasoning engines. But nudging them to the desired output is challenging. They might return categories outside the ones that you determined. They might return multiple categories when you only want one (or the opposite — a single category when you want multiple). Even if you steer the AI toward the correct answer, parsing the output can be difficult. Asking the LLM to output structure data works 80% of the time. But the 20% of the time that your code parses the response fails takes up 99% of your time and is unacceptable for most real-world use cases.

[0] https://twitter.com/mattrickard/status/1678603390337822722
rckrd
·3 years ago·discuss
Yep -- matt (at) matt-rickard.com
rckrd
·3 years ago·discuss
Thank you! I'm waiting to write this post (I follow Patrick Collison's advice methodology -- wait 10 years before you can accurately reflect [0]).

But here's Marc Andreessen thoughts:

> "Seek to be a double/triple/quadruple threat."

He talks about the MBA + Undergrad Engineering combo in this blog post. https://fictivekin.github.io/pmarchive-jekyll/guide_to_caree...

[0] https://patrickcollison.com/advice
rckrd
·3 years ago·discuss
It used to be self-hosted, but I recently moved the list to Substack to make it easier for readers if they have an existing Substack account.
rckrd
·3 years ago·discuss
Thank you!
rckrd
·3 years ago·discuss
https://matt-rickard.com

779 blog posts. Writing about engineering, startups, math, and AI.

Many of the posts have rich discussions on HN. You can see the top ones here: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

---

* Reflections on 10k Hours of Programming (421 points) - https://news.ycombinator.com/item?id=28086836

* Don't Use Kubernetes Yet (306 points) - https://news.ycombinator.com/item?id=31795160

* Google search's death by a thousand cuts (292 points) - https://news.ycombinator.com/item?id=36564042

* The Unreasonable Effectiveness of Makefiles (256 points) - https://news.ycombinator.com/item?id=32438616

* I Miss the Programmable Web (248 points) - https://news.ycombinator.com/item?id=32284375

* What Comes After Git? (227 points) - https://news.ycombinator.com/item?id=31984450

---

RSS Feed: https://matt-rickard.com/rss

Email list: https://matt-rickard.com/subscribe
rckrd
·3 years ago·discuss
For a less dramatic strategy with LLMs that expose the tokenizer vocabulary, you can use context-free grammars to constrain the logits according to the parser so that the LLMs only generate valid next tokens for the language.[0]

[0]https://github.com/r2d4/parserllm
rckrd
·3 years ago·discuss
Here's mine that runs entirely in the browser and doesn't send any data to a server[1]

[1] https://chat.matt-rickard.com
rckrd
·3 years ago·discuss
There's probably a better API that wraps generate, but there's a bit more work than the logit mask.

You have to go one token at a time, otherwise the masking becomes combinatoric rather than linear (two tokens at a time -- need to generate all two token pairs, etc.).

But otherwise, that's what the code does! https://github.com/r2d4/rellm/blob/main/rellm/rellm.py#L21
rckrd
·3 years ago·discuss
(author here) That's interesting! Maybe there's a way to quantify the cumulative probability of the squashed tokens (i.e., if you constrain to 'true' and 'false', what's the distribution of the other tokens).

For now, this is a good way to make sure that I can parse the output reliably in the minimal amount of completions (instead of looping until conformant).
rckrd
·3 years ago·discuss
Similar strategies with the logitsprocesor. It is a more generalized version that's not just constrained to JSON parsing, but any regex. JSONformer/clownfish try to parse the types syntactically.

A regex is a better fit for a different class of problems. You might implement a JSONformer/clownfish with this instead.
rckrd
·4 years ago·discuss
Maybe images are the universal interface. With some of the advancements in ML, we have different decoders: image-to-text (OCR), layout information (object recognition), and other metadata (formatting, fonts, etc.).

Now, with diffusion-based models like Stable Diffusion and DALL-E, we have an encoder – text-to-image.

Natural analogy to how humans perceive the world and how we've designed our own human-computer interfaces.

[0] https://matt-rickard.com/screenshots-as-the-universal-api [1] https://twitter.com/mattrickard/status/1577321709350268928
rckrd
·4 years ago·discuss
Large fixed costs, economies of scale, whole product (need compute AND storage). It's possible that a company like Cloudflare can disrupt certain verticals in storage that are mispriced by AWS yet have a larger than expected TAM.