I'm trying to find or invent a job + source of income I'm happy with, likely switching my career track. Mana bounty rules and USDC bounty rules below. Longer context about my skills and career follows.

You can manipulate this market to YES by helping me. The market will only resolve if I land a role I'm happy with, AND it's a direct consequence of manifolders helping me. (I. e. if no one makes any helpful suggestions, but I find or create a job I'm happy with through my own effort entirely, this still resolves NO.)

(It also counts as a YES if I get a personal research grant, create, with your helpful advice, some sort of project of my own that brings me income instead of being employed by someone else, etc.)

You can earn mana bounties. Cool useful suggestions in the comments for any avenues I haven't thought of exploring will be rewarded with mana. This doesn't mean spamming me with a lot of different potential employers if I don't feel they're relevant; but it might apply if you offer interesting career tracks I haven't realized exist, or creative attempts to unblock me ("Have you tried making a Manifold market with bounties?" would have been one), and yes, suggesting an employer I haven't heard of if they're a great fit. This is entirely at my discretion and subjective, sorry. Basically brainstorming options that would make me go "wow!".

(Even if a comment is not rewarded with mana, it may still bring a YES resolution closer, and create a useful discussion etc.)

If someone makes a great contribution that's clearly outstanding among others AND results in the market resolving YES (e. g. makes some warm intros to a company which ends up being a great fit, brainstorms my ideas with me to such a degree of usefulness that it ends with some of them bringing me my next source of income, etc.), I may decide, at my own discretion, to send this person up to $100 in crypto, in USDC. (If the person is willing to share their on-chain address publicly, I will provide the transaction hash as proof of payment also publicly). If there are multiple people who contributed together, I may split the bounty among them. Or, in ambiguous cases, send the money to the person who made the contribution first. This will be subjective, at my discretion, not subject to appeal, etc. The bounty may not be awarded if no worthy candidate is present.

NB: I'm writing under a throwaway account, because one of my coworkers follows me on Manifold, and I'm not yet prepared to announce my decision to leave. But I've been an active Manifold user for several years; I can dm you my real account, and send a confirmation from the real account as well, if you so request. You may check my profile to decide if I'm trustworthy.

You may ask me questions in the comments, which I might answer or decline.

I intend to try to be transparent with my progress in the comments by posting updates with some frequency, though it may not be very frequent.

[LONGER CONTEXT ABOUT MY CAREER AND EXPECTATIONS]

I am the only ML engineer in an ML startup. (More specifically, I do natural language processing, and the startup is in the field of machine translation).

My role is not in training LLMs; we are using smaller, more lean, more affordable solutions suitable for a startup. I've been enjoying the zero-to-one role of an early employee, creating the relevant infrastructure etc. But as the startup became more formed and advanced (and I think rigid), I've recently become less comfortable:

· I've recently developed a severe sensory sensitivity. Working from home, when it's available (e. g. when the office is on repairs), is life-saving for me, but my management is not willing to consider more frequent remote work for me. Adaptations are less than perfectly helpful. My energy is drained at work and I'm barely able to survive it.

· I have lots of ideas about the world which I would like to explore, but since I'm extremely low-energy due to a constant sensory overload, they almost don't advance. I see myself as a creative person, and working on the same thing for years does not appear very healthy for me.

So, the minimal requirements for my new source of income are:

· Remoteness of work, freedom of location, ability to create my own environment (in a dark room, without noise, without commute);

· Being able, to a significant degree, to pursue my creative ideas, instead of executing someone else's vision;

· Not actively harming anyone (I regard "progressing AI capabilities" as "harming", such is this market, sorry).

· Not just pure software development in an AI company;

· Not less than what I currently make (about $65000 a year after taxes).

Now for my skills. They are eclectic.

· Initially I studied linguistics, and was in love with languages — it's still useful, for example because I can read our datasets in many languages and see what goes wrong where.

· Then I switched to ML (because I fell in love with natural language processing), and did a PhD in controllable language generation (didn't finish it).

· Our linguistics BA had lots of math classes.

· Last year, when I came up with an idea for a dApp that I thought would be good for civilization to have, I started to figure out web3 development and created said dApp myself despite having to previous exposure to web3 development.

· I'm quite good with Python.

· I have a mild crush on molecular biology, though it doesn't quite like me back.

· I was frequenting (and sort of even helped run) a reading group on AI safety and interpretability; I think I have a minimal grasp of the field. AI interpretability (especially Anthropic-style mechanistic interpretability, though it's kind of out of fashion) is the only part of AI that still feels interesting for me.

One obvious way is to find an employer who finds something valuable about some cross-section of my skills (like when I was hired to the machine translation startup and could actually read and understand the data instead of just "stack more layers"). Another way would be to direct some of my pet projects or ideas in such a way that they would themselves pay my bills. Here some leads and also potential avenues:

· I have created a decentralized project called theorem marketplace, where the idea is that you can declare a bounty on a theorem and someone could collect that bounty by providing a correct formal proof of that theorem (in Lean).

My reasons for creating it are described here and here, but the idea is that I was trying to create a better, decentralized infrastructure for creation, storage and connectivity for human knowledge, and to incentivize research better.

It is not currently set up to bring me any money, and it currently it runs on play money (Sepolia Ether instead of mainnet Ethereum). One way to make money would be to start a prover when people submit more bounties. Another is to consider making a broader project for decentralized bounties not just for theorems; e. g. what about bounties on molecules or something?

· I have an idea I call Bayesian linguistics in my head, with an initial version described here: (I owe people an update, though), where I propose that we should model the emergence of language as a means of exchanging arbitrary Solomonoff-style hypotheses. The only way I see to make money off of this is either by a research grant or by tying it, for example, with AI safety (as a way to model and potentially decipher neuralese in which two AI agents might converse) and working on it in some lab.

· DeepMind recently announced a product called AlphaEvolve, a generalized algorithm for improving algorithms. They used it for, like, improving the algorithm of multiplication over 4x4 complex-valued matrices, and many such things. But I suspect it would also be possible to use it to actually work on proofs (from having a proof that consists of Lean's placeholder sorry toward a more and more well-formed and structured thing until the entire body is ready). Unfortunately, they closed their "expression of interest" form, so I can't get access. I'm asking my friends if anyone can get me a warm intro to the anyone in the AlphaEvolve team (I will give you mana bounty if you can). I don't see a way to make money off of this project, it would be simply about making a cool project and making some potential collaborators notice me.

· Just researching the grant landscape for any of my pet projects might be an idea.

· I would also consider working in AI safety, although if I don't come with my own agenda, I worry it might come to another situation of just executing someone else's vision. When brainstorming with LLMs, one place they mentioned was FAR AI, who specifically mention they welcome first-time research scientists and wildcard ideas etc., so I'm wondering if I should come to them and offer to study the emergence of communication to decipher neuralese to ensure talks between AIs are understood...

Another piece of context: by the law of the country I currently reside in, one is to hand in a resignation notice a month in advance, but I suspect I would warn my employer several months in advance to ensure a smooth transition, so that they find a good replacement and train them etc.; I don't want to mess up my stock options.

• Update 2025-07-05 (PST) (AI summary of creator comment): The creator has reiterated that if they find a role through their own actions, such as approaching a pre-existing personal contact, and not as a direct result of help from Manifold, the market will resolve NO.