ADA will never again trade below $0.40.

Yet a journalist may title an article like "scientists don't understand how balls work". This is a fundamental problem with media representation of AI today. We know how LLMs work. The results simply are non-deterministic, aka unpredictable.

Ask a random person on the street to throw a ball. With only these instructions, will you know exactly where the ball will land?. We know how humans throw balls. And we know how balls move. But the final result is unpredictable.

AGI needs a data-less model. Like how a child learns without being connected to the internet.

LLMs won't evolve into AGI. A use case for AGI will be to create scientific breakthroughs. Science requires creating hypothesis based on incomplete data. LLMs require big data. They break when data is incomplete.

Job market is unhinged. tbf I also wouldn't entertain any offer lower than $200k cash. But we're in polarized world where some people can't find any job, and others are declining offers many would beg for. Most these AI companies make no profit either. Eventually the music stops.

Not true, there was the Tea app launched recently.

This is actually interesting. AI without LLM. May be a shot at real AI.

I conceptualized a delta neutral stablecoin for Cardano in 2021. At the time, Cardano lacked the DeFi primitives to make it a reality. I think today Cardano is still lacking without making serious compromises. Potentially can use Strike. Realistically will probably use a CEX.

1.3M views and a ton of likes and comments agreeing. On social media you can just say things, even if you have zero clue what you're talking about. Say something outlandish, attach a narrative, and say Hyperliquid ftw 🚀. Ignore the fact that ICOs are actually worse.

As we build out a financial operating system, it's important to think about new threats, especially as quantum computing advances. New research suggests that there may be ways to make blockchains like Cardano and Solana resilient against quantum attacks.

Even better than aggregators could be intent based systems like @_mynth_. The need for an aggregator is eliminated if you can place a request and have the best outcome executed.

Media tells us we're in a population collapse and it's the biggest threat to humanity.

I used to hire unpaid interns until I learned that they're too expensive. Takes time and resources while contributing worse than nothing. If you want experience, go pick an open source project and contribute. After you've proven yourself, apply for a paid position.

Personally, I've yet to come across a use case for an AI agent. LLMs are non-deterministic, which severely limits their usefulness in agenic scenarios. Whenever I want automation, I tend to rely on decades-old regular automation tooling that's reliable and deterministic.

Whenever I get a missed phone call I question what year it is. There's a Russian who birthed over 100 children due to inventing a chat app, AI is automating jobs, and yet some businesses are still deciding to phone people?? Crazy.

RT @_mynth_: New website landed at New frontend launched. New partnerships underway.

This could become a whole lot more common thanks to the rise of vibe coding paired with laws forcing more companies to verify IDs.

For math problems, most of the time I use a calculator. On occasion, I do the math manually even though I know the calculator will do it faster and with more accuracy. This is to maintain training for the brain. Same should be done for AI. Some days, you should limit AI usage.

Other thing is, some of the most successful projects spend nothing on marketing. They focused on product instead. Good products have a strong habit of marketing themselves. Only after gaining traction does it make sense to spend money on marketing.

I know this is a joke, but it's also not far from the truth. Community keeps complaining they get rugged. Yet marketing positions are highly compensated over product positions because this produces ROI. It's almost disadvantageous to spend money on building.