For the last 2 years Iāve been obsessed with solving one problem in AI:
Reliability.
How do we build systems that donāt just generate plausible languageā¦
but can actually reason through real-world constraints and return grounded, correct outcomes?
I believe the missing piece is architecture.
Today Iām sharing the Tama Engine.
Maybe all this ai doom narrative is to offset people back to being pro ai.
1. They tried shoving it down everyoneās throat in the beginning when it was not very useful.
2. People were skeptical and started going against using it.
3. Now ai is actually useful so to counter the negative sentiment they themselves start the reverse psychology narrative.
4. People go in the opposite direction and are now defending ai progress.
I do think the next generation of coding agents or any agent will not need a fixed "system prompt" right now they feel like static pages, from back in the days. The point of having AGI is so that software should fit the person using it. Having a 'one size fits all' prompt goes against the idea of AGI. If you believe in this, follow me, I'm cooking something for you.
The problem with MCPs is that if you have 20 services you want to connect to, each of those services have 20 tools all the tools get added to your context, which adds to the cost of your agent. It consumes your allowance etc... Most services i've noticed have over 20 tools so you can imagine things get pretty bad with token usage. IT does not have to be this way though.
I believe I have solved this. More details to be revealed soon. It doesn't require inventing a new protocol.
š
Imagine if you had 20 services and each of those services only need to exponse 1 tool. Your MCP would just have 1 simple tool to call. so you can add 100s of services and it still wouldn't consume your context window because each of those tool definition consumes the same about of token which is like 100 tokens
Imagine being able to add 20 services with 100 tools each but having your agent only having to add 100 tokens to your context window for those services.
Zack Siri retweeted
Two economists mathematically proved that AI will destroy the economy.
Researchers from Wharton and Boston University published a terryfiying paper called "The AI Layoff Trap."
They mapped out the economic end-game of the AI transition, and it exposes a fatal flaw in competitive capitalism.
When a company replaces a worker with AI, it captures 100% of the wage savings.
But that displaced worker is also a consumer. When they lose their job, they stop buying things.
The company gets all the savings, but the loss of consumer demand is spread across the entire economy.
If there are 20 competitors in a market, a CEO only absorbs 1/20th of the economic damage their layoffs just created.
So every single rational CEO has a mathematical incentive to automate as fast as possible.
They can literally see the cliff approaching, and they still step on the gas.
It triggers an unavoidable Prisonerās Dilemma. If you don't automate, your competitors will, and they will crush you on price.
It doesn't just hurt workers. It destroys the businesses, too.
The economy gets trapped in an automation arms race. Companies fire their workforce to stay competitive, until the entire consumer base is completely hollowed out.
At the limit, the paper concludes: āFirms automate their way to boundless productivity and zero demand.ā
And the scariest part?
The researchers mathematically tested every popular fix.
Universal Basic Income? Fails. It raises the living standard but doesn't change the corporate incentive to cut jobs. Retraining? Fails. Worker equity? Fails.
The paper proves that more competition actually makes the collapse happen faster. And "better" AI makes the damage worse.
The only thing that mathematically stops the collapse is a targeted automation tax, forcing companies to pay for the purchasing power they destroy before they automate the job.
I have an observation. People around me praise Chinese for their developments and AI models yet when you ask these people what they actually use on a day to day basis I always hear either OpenAI or Claude from time to time I will hear Gemini and Grok. I never hear Deepseek or Qwen or Zai being mentioned.
When running agents you need to be able to see what it does. Guide it, and tell it where it can improve.
This is how you can review the thought of an agent. It's like peeking into someone's brain. Then you can steer it by adding comments.
Comments are aggregated for the next loop.
Hey everyone, I'm looking for a Senior Elixir Role, if anyone has an opening please let me know. I've built many things with Elixir / Phoenix stack including core banking systems and AI powered systems. I can bring my expertise as a leadership role or as an individual contributor. If you have an opening please do reach out!
You can also check out my profile page to see my latest app. I'm also available on LinkedIn. Please DM me if you require any further details about the work I've done.
Platforms make it harder for small companies to get distribution, so fewer small companies grow into employers. As fewer independents break through, large platforms consolidate more power, automate more, lay people off, and make the system even harder for the next wave of small companies. Users keep feeding these platforms because it feels convenient, not realizing they are helping destroy the independent internet that would have given them more jobs, more choices, and more leverage.
The internet has become a self-reinforcing employment crisis.
š



