Christopher Altmann

Information Market Engineer | Helping companies and protocol teams with Prediction Market Review, Coverage Analysis and Event Contract Design

[Numa @_0numa]

When a prediction market says "60% chance" — does YES actually happen 60% of the time?

I tested this across 113,823 resolved markets on @trylimitless.

The answer blew me away. 🧵

[Christopher Altmann @Chralt98]

There is a power in making yourself known as the most stupid person in the room, be okay with it, and then ask the questions everyone seems to have already assumed the answer to.

A lot of this has to do with you caring about what others think. If you don't, you are free to ask.

[Christopher Altmann @Chralt98]

It seems like, with all the AI-spam, the texts get longer and longer to be precise as possible, but I appreciate authentic messages with flaws nowadays.

[Christopher Altmann @Chralt98]

Human imperfection sells in a world of polished AI output.

[Christopher Altmann @Chralt98]

There is so much complexity in designing governance systems that are trusted.

I read about ZK-STARKs and find it fascinating and see the urgent need to implement it in open source protocols for on-chain governance mechanisms, especially a trusted Futarchy solution.

[Christopher Altmann @Chralt98]

The more you know, the lesser you talk, because it's often exhausting to explain it.

[Christopher Altmann @Chralt98]

The most productive time right now is to train your personal OpenClaw system to your everyday needs.

[Christopher Altmann @Chralt98]

"The verdict is consistent: token-weighted voting concentrates power, delegation doesn't fully solve it, and fairness requires more than smart contract design."

https://t.co/UcuNvyIP3z

Lets learn and use information markets and a governance layer on top of it.

[Christopher Altmann @Chralt98]

The human-in-the-loop is the necessity to prevent the unnecessity.

[Christopher Altmann @Chralt98]

It's a skill to review the plan output and reject non-necessary suggestions. Instead of the tutorial hell, there is a suggestion hell; a suggestion loop in which one constantly generates and gets exhausted by the sheer amount to review. I started to ignore more LLM suggestions.

[Christopher Altmann @Chralt98]

AI is unstructured, non-deterministic. The business of the next years is to bring the fuzzy output of AI into multi-step workflows to solve the worlds problems. A imagine solutions to problems using a secret sequence of structured prompts instead of DIY prompt exploration.

[Christopher Altmann @Chralt98]

B2C? No. B2B? No. B2A. Business to Agent.

Just an API, a CLI and open documentation.

[Christopher Altmann @Chralt98]

It's so frustrating to see how people try to play with the old book of rules in a completely changed game. Just move where the market moves.

[Christopher Altmann @Chralt98]

It's more important now than ever to write the specifications in Markdown first.

The amount of code generated by AI is unprecedented.

[Christopher Altmann @Chralt98]

Firstly, start with finding a niche.

Secondly, look for problems in search of a solution.

Then write a step-by-step process to solve a problem with multiple Markdown files.

Go deep and describe exactly how you want it.

Lastly, automate all the steps with an agents framework.

[Christopher Altmann @Chralt98]

Voldemort from Harry Potter knew a lot about dark energy, but had blind spots in other important areas of magic.

Just make sure you don't end up like Voldemort.

Move from idealism to realism by constantly learning through a variety of books outside of your usual path.

[Christopher Altmann @Chralt98]

@logansaether I end up writing long .md files to describe exactly what I want.

[Christopher Altmann @Chralt98]

AI is fuzzy. It needs continuous feedback about the expected result. Needed is a clear set of rules with weights, trade-offs and nuances, which are again defined by humans in computer code. Code is more reliable and therefore trusted. That's why I think coding will persist.

[Christopher Altmann @Chralt98]

To really become non-delusional, non-idealistic, grounded in reality is the best way to improve ones life. For this, constantly learning in many areas of life is the only solution. Thankfully, we have AI now to detect lies, misinfo, but even AI could rely on false data fed to it.

[Bill Laboon | Web3 Foundation @BillLaboon]

I have always viewed blockchain technology as an amazing truth machine. It lets anyone trustlessly and mathematically verify that something took place at some point in time - really remarkable!

This truth machine is even more important with recent improvements in AI, which in some sense is a sort of "anti-truth machine" - LLMs tend to hallucinate, generative AI lets you create anything you can describe, etc.