brainless

A year ago, it was not sure if the core companies around programming would move to more agent enabled workflow. Companies like JetBrains.

I love their products. But it's clear now that programming is now spec management and how do we control agents. I don't need JetBrains for that.

And spec management will become less technical as models get better. What does JetBrains think is the long term goal for this product? I don't think it'll make any sense in a few months.

Forgive the skepticism: Is this a harness that sits beside a language runtime, looks at traces/logs and uses an LLM to optimize code for given hardware?

It looks interesting but I am really unable to see what is so different from a couple days of prompts and digging into a software to optimize it using LLMs.

I do not think OAI/Anthropic make the biggest share of their revenue from subscription. The largest share is from API spends. That is what DeepSeek or others aim for as well.

This is lovely. Tried on speaker, will try with headphones.

I have a chicken farm (and a pig farm but I manage the chickens myself). I want to start a duck farm with a fish pond. Living in a small village and enjoying these things in real life actually changed my life in ways that I cannot explain.

I still (get agents to) write software but my daily life is around the farm and village life (mostly a Sun centric routine). I have to share a bunch of photos, not doing that enough. And I want to record all the bird sounds. Perhaps host a livestream. I live here; the home stay is by the host family: https://maps.app.goo.gl/TEUSdwvko2DRWmou5

Very similar boat. In my twenties I used to look down on working with hands - like carpenter, farmer, plumber. Now, I want to only do this. I switch between farm and desk work all day.

I started bake a couple years back but I am just not consistent. Need to get to that, would love to have fresh bread in the morning since we don't have that in any village nearby.

I am 43 and only in the last few years have things turned better. I want to say that life is tough for different people from different parts of the world. For various reasons. But overall we have a set of systems that do not care.

The lack of friendship, care, love cannot be eased with money. I used to be crazy about startups. I used to want to live in New York (I am from Kolkata). I had a great run in my twenties but I crashed many years ago. Living in Germany helped a lot. I realized I was not loved or cared for. By myself mostly.

My parents love me but my dad has mental health issues which took a really bad toll on my childhood and teen years. I swept away the need for a love and care and replaced it with money. Many of us are doing this. The system needs most of us to slave. Now, I have been living the village life for about 6 years. I moved my parents as well. We have farms. I still write software, well, mostly generate them with LLMs.

There is a sense of purpose here that I started to feel. Over these last years, I have become more calm than I ever was. Happier, in a way I notice, every day. With less money than I ever had. Life is not about lots of achievements. We need meaning, love, purpose, care, friendships. We need slows walks, flowing water, walk in the rain, healthy food with zero labels. We need to grow our crops, care for animals, raise kids as a village.

I know I am typing this on a computer. But I no longer work for any tech company. I have a small consultancy client. I live as per a typical countryside routine. Most of my (whatever little savings) money is now invested in our farms where I work a few hours everyday, on time. Animals, dirt, mud, poop, crops, cleaning. All of it. Connected to how life is supposed to be. I used to hate glass buildings, still do.

Concrete, glass and everything about artificial spaces makes me sick. I am just one person, I do not have some research paper. But I know how I feel today vs a decade ago. Hope it helps.

I remember CGI days and I have written tiny web "features" with Perl and C many years ago. All my product backends are now in Rust and SQLite but I use frameworks and libraries. Like Tokio, Auxum/Actix Web, Diesel.

I am not sure I like the idea that the binary will be called by nginx on each request. I prefer the backend binary to have state. A 256 MiB virtual instance is enough for any software shared between friends, family or even a small community.

SQLite backups are easy. I really like how I am building web apps these days. Been with PHP for 7 years and then Python for 10+ years. Single binaries are awesome. I am looking at multi-server backends as well, still 256 - 512 MiB per instance but with a shared Redis. Still with SQLite. Should easy enable medium sized business software.

To me, "self-hosted" means, either I use a rented hardware or I use a purchased hardware.

If the software can be switched to a rented or purchased hardware (I do not want Cloudflare's offering) then it is "self-host-able" IMO.

I think there are open source Cloudflare alternative hosting software.

I am not an expert, just an engineer with years of experience. I switched to Rust a couple years ago. I was still learning before LLM generated Rust was good enough that I stopped coding by hand.

Again, not a language expert, but: if you can describe your domain or business logic as clearly as a strongly typed language with a rich type system, most of your issues are gone. Enum and Struct with all the other core types plus the match expressions does most of my mental heavy lifting. I do not even track any of the recent language changes.

What I really care about is the shape of what I am describing - does it translate to code? How much do I lose in the translation? I want to try other languages, particularly Lisp but Rust is at this moment my choice. I have my own UI framework (1), my own provenance based business domain generator and a few simple language parsers.

I am building apps where you can, for example, throw a CSV file (2), ask questions in English and get answers - without using an LLM. Parser. Rust is no doubt a great language to express - not as a programmer, but as a prompter. I do not write the code. I ask LLMs to generate it using only basic knowledge of Rust and its type system. This will be the key to work with LLMs for majority of people.

1. https://github.com/brainless/akar

2. https://github.com/brainless/baho

Thanks again, I have hope.

SQLite/Turso fits my own agent/harness mental model very well but I am, as folks in the valley would say, a bit of an anti-capitalist.

I am working on my own agents that are anti-tokenmaxxers. SQLite/Turso is part of the deployed app architecture from my harness. Thanks a lot for taking the time.

Huh? Supabase is also that. Lots of tech is like this. Parent company makes money from hosting, services, enterprise feature seats, etc.

You mean, it has never happened that an acquiring company said this and then closed the acquired tech?

I am fully onboard with LLM generated code if that is what you mean. I literally get client work titled "Need Claude led Engineer". SQLite is a great option for the LLM enabled future.

Thanks but you do understand this makes it hard for folks to select Turso. With all good intentions parent companies kill acquisitions.

Supabase is hosted Postgres with wings. Turso seemed like it would be the same but using SQLite based approach.

So either Supabase now uses the Turso sync + Postgres or something that needs Turso or abandons it. Not sure and I am not asking, just sharing my thoughts.

Love what you folks have been doing.

I do not feel good about this. I really want SQLite (and the Turso re-write of it) to win for tiny apps that people will build more and more with LLMs. Desktop or server. Even serverless.

The core tech is open source so that is good. Will it stay? I guess Turso hosting was not making enough revenue?

What I mean is that a lot of the popular languages have an engine written in C. Fairly common pattern. I get it that we want to abstract but sometimes it feels we just want to avoid dealing with a language that is one layer below.

Now, LLMs can.

How many of the popular languages are running C underneath? Count them, I will wait.

If a language is Turing complete, why build another language right on top? Most of the interpreted languages are running an engine written in C. Was C not enough? Or was it that we needed abstraction?

It is nice to be in our bubble but if what you said is true, why do we only have like 3 major instruction sets? Why not 40? Programming languages are just easy to create - this is a good thing and we created too many. But it absolutely fractured the industry.

Look at the average job post. There are requirements for system design - good. Algorightms - good. Then it goes into this whole language/framework land. Remembering syntax is literally there in many companies interview. These are preferences. If we understand how a computer executes binary/assembly, the layers above are not as relevant as we have made them.

I have personally wanted to learn and actually use Lisp. I will get to it, even with all the LLM enabled programming. But yes, this is one of those things.

We have reinvented so many ideas many different ways in many different languages and frameworks simply because we want our flavored version.

I have been programming for about 30 years (including school years). Professionally for 18 years.

Can anyone tell me why we have 40 or more programming languages, with about 10 popular ones? Then about 20 frameworks in each of them. And add another 200 popular libraries for each language? This matrix make no sense till you realize - it is preferences all the way down.

Most of us engineers have built our own mental model of programming. We are all right. But the users do not care. LLMs are here to produce code closer and closer to the metal as needed. They can sit and create a graph out of every spec, use an AST that they develop and run on the CPU if they have to. They will do it. No amount of us discussing will stop that.

Programming is going to be re-invented. I do not think the current ways to write software will even matter.

There are lots of distilled models on huggingface that are much smaller than, say, Opus. They are distilled from Opus or Fable and show clear improvements. I do not have the budget to fine-tune a 30b or more parameter model but from my tiny model experiments, the results are quite clear. Again, I have only a couple small tests.

Have you actually fine-tuned yourself? Email categorization comes to mind and there are tons of non-LLM approaches even that will give fantastic results. How did spam filters work before LLM?

I think LLMs just made us think that is the only way. It is not.

You can go quite far using a human language to Bash grammar based setup but at some point the input prompts are harder to translate. The OP has existing projects that work with AST quite deeply so I assume they know about that already.

I am building a natural language to CSV/Excel commands for a "wrangler" type desktop app. Same issues. The MVP is being built with parsers of sorts, entirely code generated. Then I want to fine-tune a tiny model at some point.

https://github.com/brainless/baho

I have been trying a mix of fine-tuning and I am amazed that most people do not see this coming.

A tiny, smaller than 1b parameter model, fine-tuned, can kick ass for constrained work. I do not have a lot of budget, I fine-tune only on a 16GB M4 Mac Mini. But that also tells me the potential is wild. Progress has been slow since I moonlight on this.

I have been trying to build a set of models + agents for full-stack development, where each model does only a small piece, like take user prompt and break into backend/frontend tasks. Then a Rust+Diesel model, a Rust+Auxum model, a Solid+Router model and so on. I know this is wild but this is just theory - can 5 or 6 Qwen 3.5 0.8b models do full-stack web development? My hunch says they can, better than what most people expect. Heck, with a good harness, it might beat all the cheaper models for the specific task, like Haiku or Luna.

I think LLMs may actually help us get to CLI/API driven development. At least, that has been my experience.

Even though most of my projects have a UI, I build a CLI/API version so that the LLM can interact with it directly. I have been using this "CLI driven development" approach for more than a year now and have had fantastic results. The CLI arguments make it easy for LLM to interact with software it wrote.

I usually ask LLM to build a lib, then expose as a CLI and a RESTful API.

I came across this recently. I was scanning for tiny models from HF using their search API. The script was generated by an agent. When I ran it, Qwen 3.5 did not make it at the top. Turns out, models generally prefer older content (training) but that the scanner also did not give any importance to recency.

I was learning Rust slowly when the LLM enabled coding became good enough. I switched from learning to full on building with Rust. I still learn high level concepts as needed but I will not be able to write Rust on my own at all.

And that sounds scary but the way I got over the fear is by realizing there are many things that I do very well but I do not know their internals very well. Driving is an example. I barely understand what the steering wheel, clutch or brake pedals do. I have driven over 130,000 Kms and I will perhaps drive more than double that in the next many years.

I have been building software since PHP/Drupal days. Got into AWS S3 as a beta user. Adopted Memcached (and MQ) in 2008 out of necessity. Then Python/Django for 10 years. Then Rust. And tons of JS/TS. I owe a lot to my curiosity. I believe we can keep learning what we need and still delegate most of programming to agents.

I am not an expert in this domain but as an engineer-turned-researcher, this looks a lot like GliNER with a fitting harness.

This is something I focus on in a bunch of my experiments - how to get immense value out of tiny models (<1b params). There are lots of different architectures out there and there is so much to optimize if you know what you are asking and have a grammar to constrain with.

Great to see this and I hope this is a lot on top of what is already openly available.

Yes they are great for learning at own pace, trying out new things.

I have accepted two things that make me a happy engineer now: AGI is not here no matter what they say and LLMs are still very useful if one knows how to use them.

They are another layer of abstraction and like you said they do not tire. There is a lot of optimization needed so we can reduce wastage (running 1T+ LLMs for most work is wastage).

More and more such experiments. I felt sad for a couple months when I realized that writing code will not be the same since. Now I am on the other side.

LLMs are interesting in their own ways but as an engineer, this is a way to unlock a new way of building software.

I recently build a Claude-assisted Excel/CSV parser for a US based property management system (tax compliance). Uses Haiku and has a lot of deterministic code to extract column/row combinations to check known formats and finally handing out the headers to Haiku to give us a translation plan to our support columns.

These would eventually become part of the software, in a tiny LLM. The gap between training (such tiny LLMs) and inference will shrink. We can consult Claude for edge cases, create sample dataset and train a the tiny LLM on demand so we go to Claude less.

The tooling that a project needs is really important. Something I have been feeling as well. Not just in LLM building projects, but regular software projects that are LLM generated.

When we hyper focus on finding something, we find it all the time.

The world of tiny LLMs is so interesting. It is unlocking novel ways to encode information. Why focus on the style of writing instead of the subject matter?