Would be curious to hear any feedback from folks using Grain.
I've been following a similar WASM-first language, Moonbit (https://www.moonbitlang.com/), which was created by the creator of Rescript. I'm excited by the design decisions they've made and think there is a lot of promise on the AI story for these fully typed "low resource" functional languages due to their expressiveness. My bet is that coding agents will start to really benefit from the type inference they provide.
Hiring: CTO (to take reins from solo, technical founder)
PHOSPHOR (http://phosphor.co/20x) has been reverse-engineering ISDA, which financial historians refer to as “the greatest success story in modern economic history” due to ISDA's ability to drive standardization into financial contracts which enabled their markets to grow 1200x.
We're doing this in the internet age by modeling Github as the center of the open source world. But to build something like Github, we had to build languages, version control, IDEs, and infrastructure for contracts and financial models (the backbones of every asset class in finance) to make them computable and connectable.
I've been building Phosphor for over three years as both CEO and CTO. As we've really honed in on what we're doing as a business, how/where we innovate, etc, I need someone experienced to take the reins and help me grow this company into the ridiculous opportunity it represents.
You might be a fit if:
- You're demonstrably obsessed with HCI and end-user computing
- You’ve lost sleep thinking about knowledge abstraction problems
- You’ve built CRDTs and collaborative applications for fun
- You're design-oriented, but as speed is a UX problem, you’ve learned systems languages
- You’re excited by the emergence that occurs when AI is applied against observable systems
- You're excited to
- You’re a polyglot, and maybe frustrated that functional languages aren’t the norm
- You have an expansive model of the world, and the idea of consolidating these skills to create extraordinary real-world impact excites you
- You're deeply technical, but have also built capable teams
What you’ll do:
- Build one of the most unique products in finance
- Inherit a three year old, thoughtfully architected, event-driven application with only minor tech debt
- Build a 5-10 person engineering team
- Kick off/manage all of the bits and pieces that go into such a ridiculous endeavor
This is a very hard technical problem at every level (low level up through design) which we've made extraordinary progress towards.
We have traction and investment from leaders in clean energy finance (where I built one of the leading solar platforms as first hire), real estate, and more.
To apply, email oliver@phosphor.co with "CTO position" in the title. Include a PDF of your resume, links to LinkedIn/GH, and lay some bullets that point to awesome you've done or built before that match the bits and pieces of the "You might be a fit if" section.
This is very very very interesting work. I had a similar feeling and ideation about Law and legislation before, and envisioned a World whee laws could be written in a syntactically structured language and then formally verified to fix, prevent and elaborate new rulings, eliminating ambiguity and creating the same type of standardization you describe.
Then I got severely discouraged by every single lawyer friend I have as an impossible task, both technically (which I disagree, it still sounds perfectly reasonable to me) and most importantly: politically. There would be no interest from real world actors in this, because all of them profit exactly on the ambiguity of the law.
I'm genuinely curious to understand why that would not be the case here, and ISDA's case seem to be a good inspiration.
We've had "fun" solving this for Phosphor which has similar parts to Causal, but is focused more on real assets and emphasizes connection with computable contracts.
For circularity, we found that we could keep the UX dynamic by making a deep copy of the circular part of the DAG behind the scenes, asking the user to determine which variable in that path should be "resolved", hard coding that variable in the copy, then solving that variable to zero through a newton optimization. Once optimized (in parallel to main graph), it feeds back into the main DAG just like any other dependency.
Would be a silly approach in Excel, but not so much here.
> Use SQLite for everything, or Postgres if you outgrow it.
Or both! I've followed ElectricSQL for a while and it's pretty awesome to see how they've approached it. Rest APIs replaced with client side SQLlite which syncs to server side postgres over a CRDT layer that you don't actually have to touch/know anything about. And they just rebuilt postgres on pure WASM (https://github.com/electric-sql/pglite).
It's just cool to think that maybe you don't have to do design all of these intricate REST apis and just call SQL.
Not yet, it's a bit rough. The LLM I am using requires a bit of extra fine-tuning to be really smooth, I need to rent a bigger GPU. Besides, I am working on some novel integration between transformers and SAT/SMT that will take me some time to finish.
Is the theory tied to a specific llm? I'm interpreting it as, e.g., the llm writes the code, the solver verifies it, repeat until correct. In this situation the two are decoupled and the llm would be a drop in and thusly could be any local or remote llm. Is there something about your approach that doesn't allow this?
Our product (phosphor) is built end-to-end in F# so this stuff is close to my heart. You might find Moonbit's approach to functional AI interesting as well https://www.moonbitlang.com/blog/moonbit-ai.
For what it's worth, Moonbit (https://www.moonbitlang.com/) is a really nice take on this. Designed by the guys who created Rescript for OCAML, but for WASM-first world.
After reading James Nestor's Breath, I went through the Oxygen Advantage certification for fun. Lots of studies that show that the 4/6 or 5/5 cadence (ten second round trip, leading to six breaths a minute) strongly improves heart rate variability and engages the parasympathetic nervous system.
OP had a cool idea -- wish it hit the 5/5 cadence.
PHOSPHOR (http://phosphor.co/20x) has solved the last "impossible" technology problem in capital markets.
We've represented financial models AND contracts as structured, computable data, and we've done so without sacrificing flexibility afforded by Excel or Word. This has never been done, and it enables product opportunities that have never been conceivable before: we have the first real view towards AI that does not hallucinate in the most business critical mediums in finance.
I founded Phosphor after building one of the leading clean energy investment platforms in the US as first hire. Phosphor was originally a clean energy finance product, but then we solved a much deeper set of problems. Over three years, I've led three senior engineers as CEO/CTO while we've worked (in stealth) through mind-boggling R&D problems in computable contracts and linguistics.
As we focus on commercialization & launch over the next few months, I'm seeking senior product/UI engineers, and I'm beginning a CTO search in hopes of finding a senior technical leader to take the reigns and partner with me to build something OpenAI-scale.
For the adventurous of you, resumes to oliver@phosphor.co. Please mention HN in the subject line.
