Death to living in ignorance
This is no time to let life happen to you.
A quick word from the author:
Bongiorno,
While hanging out with my house-mates around the fire yesterday, we started talking about the reality that surrounds us and current social, political and economic context.
We all had different takes, some of us mentioned this is the best humanity has ever been and we’re lucky to be alive today, some of us (myself included) thought that this is just the beginning of the end and we’re doomed to experience techno-feudalism.
There is an objective reality and a way to interpret that reality (our own subjective views).
The objective reality is that inequality is getting worse every single minute, government is powerless (or actively trying to make it worse) and we’ll soon all be at the mercy of the few and the powerful.
Now, taking into account this objective reality, what’s our own individual take? There’s a lot of options, for example, we could:
A. Decide the bad guy in this current setup are the corporations, the elites and the 0.1% and fight them through any means necessary (backing political adversaries, boycotting the companies owned by these elites, speaking out and protesting against them etc.)
B. “If you can’t beat them, join them”. Start building products and services for the elite, try to join their ranks by making things they want to see in the world. Praise them on social media for their skill and know-how and support their actions.
C. Build your own, parallel system. Disconnect from the current architecture of society and find alternative ways to live and grow (this is the hardest to put into practice since our world is so inter-connected). This is reflected in what the crypto community initially tried to achieve.
D. Live ignorant to all these struggles and just focus on your own well-being and your immediate family. Get a job (any job at this point), go on vacations, enjoy life to the fullest and navigate the uncertainty one turn at a time.
E. Actively pursue your own narrow-goals, not focused on the bigger picture. Build that restaurant/club/app you’ve always wanted to make. Very similar to option D, but with more intent on personal goals.
F. (My Take) Understand what’s under my control, what’s not, and make an active, genuine effort to move things in what I consider to be the “right direction”. Problem with this approach is that what I consider “right” will be considered “wrong” by others.
Day-to-day, I do a lot of reading of technical, professional and, let’s call it “philosophical” reading. For example, this great read from Dan Koe about the 12 rules to change your life.
The principles are incredible and will definitely help you engage with your life better than before. But I can’t escape the feeling they’ve been written from the myopic lens of a person who owes a lot of their achievements to luck, timing and positioning.
Here’s what they don’t tell you about success:
Once you’re at the top and start giving advice, you’re already wrong. Because the path behind you has already changed dramatically and things aren’t like they used to be, the same path to success, the same principles, beliefs and systems that made someone successful no longer apply to the new reality.
What I want you to take away from this rambling of mine is to be very aware of the lens you’re adding on top of the objective reality we’re all living through. The way to do the most damage to yourself and everyone around you is to stay ignorant to your own beliefs and values.
The bad and the good of last week:
“Hundreds or thousands of companies, and billions of dollars, are being spent on replacing human workers.
Some don’t think this is possible, but they think we need to invent some super smart model that’s better than anything we’ve ever seen.
We don’t need that. What we need is scaffolding and piping that connects the dots and brings the right context together in the right way to solve problem x or y.”
“In recent years, the unemployment rate among college graduates aged 22-27 has inched higher than jobless levels for all workers, according to the Federal Reserve Bank of New York.
This is a new development: Young degree holders have consistently experienced lower unemployment than the general public going back to at least 1990, the earliest year the data is available.”
“Jumping straight into execution feels fast, but it’s risky. We might end up solving the wrong thing, or solving it the wrong way. That’s where GPT can help: not just by generating ideas, but by helping us think more clearly about the structure of the problem itself.
There are many ways to break down a task. One of the most useful in product work is the Jobs To Be Done (JTBD) framework. Let’s see how we can use advanced prompting to apply JTBD decomposition to any task.”
“Today, we launched Eleven Music – the next step on our mission to build the most comprehensive AI audio platform in the world. With Eleven Music, businesses, creators, artists, and every single one of our users can generate studio-grade music from natural language prompts, with:
– Complete control over genre, style, and structure
– Vocals or just instrumental
– Multi-lingual, including English, Spanish, German, Japanese and more
– Edit the sound and lyrics of individual sections or the whole song”
To be fair, LLMs are quite good at writing code. They’re also reasonably good at updating code when you identify the problem to fix. They can also do all the things that real software engineers do: read the code, write and run tests, add logging, and (presumably) use a debugger.
But what they cannot do is maintain clear mental models.
LLMs get endlessly confused: they assume the code they wrote actually works; when test fail, they are left guessing as to whether to fix the code or the tests; and when it gets frustrating, they just delete the whole lot and start over.
Startup Idea of the Week: Collective LLM Teaching
You know what’s worse than AI models making mistakes?
Watching them make the same mistakes—over and over again.
That time an LLM explained how to boil water at “450°F”? Painful.
The one that confused Kafka the author with Kafka the database? Brutal.
And don’t even get us started on tax advice…
It’s not that the models aren’t smart. They just aren’t being taught by the right people.
Because who knows more about tax law than a tax lawyer? Who knows more about mechanical engineering than an actual engineer?
But right now, specialists don’t get a direct way to feed their knowledge back into the system.
What if they did?
What if training an LLM was less about abstract datasets and more about a collective tutoring session—where experts teach the AI in exchange for better access?
How It Works
🧑🏫 Experts Answer, Models Learn: Specialists in fields like law, medicine, design, or software development respond to real user questions. Their answers aren’t just helpful—they become high-quality training data.
💸 Teach-to-Earn: In return, contributors earn free or discounted access to the model. Think of it like tutoring your future AI assistant.
🔍 Quality Control Loop: Answers are peer-reviewed, scored, and reinforced before being ingested. Experts earn reputation points alongside their credits.
📚 Domain Pods: Communities of practice form around specific areas—e.g. “LLM for Radiology,” “LLM for Contract Law,” or “LLM for Game Dev.”
⚡ Continuous Improvement: Instead of waiting for big updates, the model gets smarter incrementally, with domain experts steering it.
Why This Works
LLMs are only as good as the data they’ve seen. Right now, that’s mostly internet text. Great for memes, not so great for medical nuance.
Specialists want smarter AI in their field—but they don’t want to give away their expertise for free.
This creates a neat loop:
-
Experts supply targeted, vetted knowledge.
-
The model improves in real-world domains.
-
Experts get privileged access to a sharper tool that helps them do their jobs faster.
It’s a collective upgrade cycle.
Go-to-Market Plan
🎓 Start with professions where precision matters (law, healthcare, finance).
🤝 Partner with professional associations who want AI aligned with their standards.
💡 Target indie consultants and freelancers: people who’d love free AI horsepower but can’t pay enterprise rates.
📱 Build a gamified platform where “teaching the AI” feels like contributing to Stack Overflow—but with direct rewards.
What Could Go Wrong?
❌ Experts might game the system by giving low-effort answers.
❌ Conflicting expert opinions could confuse training.
❌ The platform risks becoming just another Q&A forum.
That’s why:
– Peer review and reputation scores gate what counts as training data.
– Multiple perspectives are tagged and labeled, not discarded.
– The incentive structure is tied to model improvements, not just volume of answers.
Business Model
Freemium.
🆓 Anyone can ask and browse questions.
💎 Verified specialists can:
– Earn credits for discounted access.
– Build reputation in their domain pod.
– Access pro-tier tools (custom fine-tuning, private deployments, priority inference).
Ok, hear me out, you should try: a Trustworthy Smile
Trust isn’t just built. It’s triggered.
Research suggests we evaluate trustworthiness based on facial cues that resemble happy expressions, like an upturned mouth and slightly raised eyebrows.
They signal to us that a person is likely approachable and has good intentions, so our guard drops.
But it runs deeper than just feeling comfortable.
One study found people invested up to 42% more money in strangers who had “trustworthy” features. 🤯
Because, as research shows, feelings of trust quiet our risk radar.
When someone looks trustworthy, we’re less likely to imagine getting scammed or disappointed.
So instead of spiraling through all the scary “what ifs” on the checkout page, we click “Buy now” with confidence.
That’s why trust isn’t just a vibe—it’s the currency of business.
In just milliseconds, potential customers will judge your profile picture and decide if you’re trustworthy or not.
So if you wanna trigger trust and earn more sales?
Smile genuinely, choose a profile picture that screams “I’m trustworthy,” and use that same pic across every platform.
What’s going on for us:
Our platform is live:
Is it working perfectly? Nope.
Do we know how to improve it? Not yet.
Is this gonna be the coolest f*cking thing once it works properly? 100%.
More news coming soon, for new we keep it cryptic to increase FOMO.
Have a wonderful time these last few weeks of summer,
The guy who didn’t know how to code 8 months ago.