faxmeyourcode

This is a potentially polarizing (but very welcome imo) change to pi. I'm such a huge fan of the pi SDK, and just yesterday I was trying to wire up the pi mcp extension to be used in a project and sonnet told me about the ongoing work in the 0.99.x tags. Didn't realize the next day it would be merged! Fantastic news.

This is how AI is going to kill us all. It's not going to become superintelligent and harvest our biomass for energy - we're just going to assume it's moderately intelligent and those assumptions will lead to badly informed decisions and by the time we realize we are acting on bad information it will be far too late for anyone to do anything about it.

This analogy is exactly what made it click when I was a teenager.

Firefox desktop and mobile let you install addons like https://darkreader.org/ which work great.

Cloudflare mostly because workers and durable objects are so cheap.

This is so cool! I'm having trouble thinking of all the different use cases and potential issues with this, but I don't think anyone has done something like this before. Great work

It's already here! https://www.atlassian.com/software/rovo-dev

Nice! The idea of having a friday leetcode problem for my team has been kicked around to try and keep us on our toes when it comes to solving these style problems and preventing atrophy of this part of our brain. Might try wrapping this cli in some kind of web ui or adding it to our slackbot

It's not clear to me from the article, are they asking sol to output bounding box coordinates with some kind of structured outputs?

Anecdotal but I've seen it use python to crop, zoom, and "enhance" (fiddle with sharpness and brightness) images to read sections of handwritten census data from the 1800s. Feels like that there might just be a mismatch of capabilities when it comes to straight outputting coordinates but I bet the model is better at actually finding the answer given any tools available. Which I get is a bit of an apples and oranges situation.

I've also tried to use it to identify an old pair of glasses and it didn't stand a chance, so I do think it's not quite there yet when it comes to some vision tasks.

I built macrosforhumans.com because I had been using cronometer since 2020 on and off and was familiar with tracking macros but always hated the entry. Measurement is something that you have to do regardless but the annoyance of having to suffer through their search/recipe/diary CRUD UI was unbearable. With MFH there is no mobile or web app. it's just a phone number that you text. This reduced the friction for me significantly and I've already lost a few pounds and, after recently turning 30, hopefully a few more with more diet and exercise.

I've been building this for a while but I'm working on getting my first 100 users now which is my first foray into the marketing/sales side of things and any tips would be helpful. Also I didn't realize until later that the name, while I think is pretty cool, might be extremely poor SEO? :-)

I also think that these kinds of post-app apps are a curious idea and will become more popular, although likely (hopefully not) through sms as the communication medium. I am off this week and plan to write some stuff about it but it feels really cool to remove the terrible CRUD UIs entirely from this part of my life. I tell my friends that even if it doesn't take off I'm never shutting it down because I enjoy using it over myfitnesspal/cronometer/macrofactor etc so much.

Compaction, busting the cache, and other issues like that will lose significance when you're running at 15k tokens per second like chatjimmy. Very interesting to think about what will change in the future.

Apparently there was a gathering of harp guitarists here in North Little Rock at a guitar shop that recently opened just down the street from where I live last year: https://www.harpguitars.net/2025/11/29/hgg23-in-north-little...

I was totally unfamiliar with harp guitar but I am sad that I missed such a special gathering!

I've definitely noticed 5.6 sol being extremely trigger happy in ways other models, even 5.5, we're not. I would definitely categorize a few small incidents at work where it performed "actions a reasonable user would likely not anticipate and strongly object to." Just my anecdotal experience.

For example discussing driver upgrade and subsequent password rotation and it didn't stop and ask me if I wanted to restart the service or install the driver or anything, it immediately took action. It feels like a side effect of pushing more "agency."

I feel like another comparison worth looking at is purely cost.

Capability per dollar is something I care about:

    Opus API    $5/$25
    Sonnet API  $5/$15
    Haiku API   $1/$5

    GLM 5.2 API $1.4/$4.4

So you're really getting near opus level capability for the price of haiku.

A town I grew up in went absolutely crazy with roundabouts over the last 10 years or so. They built them everywhere. Most are pretty straightforward and they simplify traffic a ton but there were some bad design choices made.

Take this similarly confusing double-roundabout for example https://maps.app.goo.gl/VTSrSWsBGnsYyzKU6

There are accidents here almost every week and when an out of towner comes off the interstate to get some gas at Sam's Club (where I used to work) I had to try and explain to them how to get back onto the interstate...

I might try to clone this project idea, it was fun to play!

They also proposed a peanut style roundabout with a *train track going through it* that thankfully has not been built yet https://media.conwayarkansas.gov/media/documents/project/85/...

I work a lot with databases, but only a couple times have I really messed with time series stuff. I've enjoyed reading your list of tricks and it's gotten me interested in playing around with writing my own toy database to just experiment with some of these ideas.

Somebody else in the thread brought up the benefit of snapshotting a database at a point in time stores not only the state of execution but also the code, etc. That is a unique benefit I'd be interested in exploring over storing your orchestration outside of the database.

Not trying to dismiss the project - it looks like a lot of hard work has gone in and somebody has a use for it. I just come from an airflow style external orchestrator frame of mind that manages durability state in postgres but keeps the control flow out. Sorry if I came off as a bit snarky

I aggree - I'm not understanding the value of the project either if you look at the example here https://github.com/microsoft/pg_durable/blob/main/examples/i...

It's an interesting technical achievement I guess, but it's very bizarre to try and read this

    SELECT df.start(
        @> (
            ($$SELECT ... FROM demo.invoices WHERE status = 'pending'$$ |=> 'inv')
            ~> df.if_rows('inv',
                $$UPDATE ... SET status = 'processing'$$
                ~> (df.http(...) |=> 'resp')
                ~> df.if($$SELECT $r.ok$$,
                    -- classify, branch, wait for signal ...
                ),
                df.sleep(5)
            )
        ),
        'invoice-approval-pipeline'
    );

This feels like the wrong solution to an age old problem solved by the DAG schedulers like Apache Airflow for a while now.

Why would I want to store my control flow in the database and not in code? It feels strange.

Not trying to dismiss the project, I'm just not getting it yet I think.

I had a hunch that opus 4.7 hedged more than other models - and it turns out it's true

    model                 total_claims  hedged_count  hedged_pct
    claude-opus-4-7       1000          451           45.1
    sonar-pro             1000          391           39.1
    gpt-5.4               1000          277           27.7
    gemini-3-retrieval    1000          129           12.9
    gemini-3-pro          1000          60            6.0

datasette query here

https://lite.datasette.io/?csv=https%3A%2F%2Fstatic.simonwil...

An HP 50g was my calculator of choice, and the whole RPN style really rubbed off on me. Plus it had more advanced symbolic algebra capabilities than a ti83 equivalent. I enjoyed learning common lisp, scheme, racket, etc through high school and college and still am fond of them today because of this calculator.

How does it compare to mythos?

Wow, I didn't realize some RFID could reach 15 feet out - that's good to know. I naively thought you essentially had to be touching the surface of the tag.

If you treat LLMs as generic transformers, you can fine tune with a ton of examples of input output pairs. For messy input data with lots of examples already built, this is ideal.

At my day job we have experimented with fine tuned transformers for our receipt processing workflow. We take images of receipts, run them through OCR (this step might not even be necessary, but we do it at scale already anyways), and then take the OCR output text blobs and "transform" them into structured receipts with retailer, details like zip code, transaction timestamps, line items, sales taxes, sales, etc.

I trained a small LLM (mistral-7b) via SFT with 1000 (maybe 10,000? I don't remember) examples from receipts in our database from 2019. When I tested the model on receipts from 2020 it hit something like 98% accuracy.

The key that made this work so well is that we had a ton of data (potentially billions of example input/output pairs) and we could easily evaluate the correctness by unpacking the json output and comparing with our source tables.

Note that this isn't running in production, it was an experiment. There are edge cases I didn't consider, and there's a lot more to it in terms of accurately evaling, when to re-train, dealing with net new receipt types, retailers, new languages (we're doing global expansion RN so it's top of mind), general diversity of edge cases in your training data, etc.

Especially for super constrained applications. I don't care if the language model that I use for my extremely specific business domain can solve PhD math or remember the works of Shakespeare. I'd trade all of that for pure task specific accuracy.

Labeling or categorization tasks like this are the bread and butter of small fine tuned models. Especially if you need outputs in a specific json format or whatever.

I did an experiment where I did very simple SFT on Mistral 7b and it was extremely good at converting receipt images into structured json outputs and I only used 1,000 examples. The difficulty is trying to get a diverse enough set of examples, evaling, etc.

If you have great data with simple input output pairs, you should really give it a shot.

> Filip also told me that he asked Claude to continue on the even case after the odd case had been resolved. “But there after a while it seemed to get stuck. In the end, it was not even able to write and run explore programs correctly anymore, very weird. So I stopped the search.”

Interesting snippet towards the end. I wonder if they were using claude.ai or claude code. Sounds like they ran out of context and entered the "dumb zone."

Yea, I agree. The dataset is < 100MB... so duckdb can very easily handle this on an old macbook air.

https://duckdb.org/2025/05/19/the-lost-decade-of-small-data

Neel, this is really cool. How long have you been working on this, and where did you guys get inspiration from? Did you work on vlms earlier or something like that? Just curious.

Also, thanks for choosing a technical blog post for presenting this information.

I love the idea behind MyVisualRoutine as a father with a disabled kiddo, thanks for sharing.

The app is beautiful - much better than I could build - what tech is it using if you don't mind me asking? Is it flutter, react native, something else? Just want to get better at mobile dev.