Passer

Observe. Adapt. Move. Three simple words. Not a motivational slogan. A way of operating in uncertain environments. A market survival framework. The sparrow metaphor. The Passer Method.

@KaiXCreator Try giving an e-bike to someone who can't ride a bike. That's what happens.

People building stuff in Claude code without proper coding knowledge scares me.

@aryanlabde Something I need, and hope others will too 😁

vibe coders, be honest: are you building something people need or something you thought was cool?

@HowToAI_ Good shit for you @0ldskultxo

Google has published a paper that might end the transformer era. For the last 7 years, every major AI, ChatGPT, Claude, Gemini, has been built on the exact same architecture: The Transformer. But Transformers have a fatal flaw. To remember context, they have to process every single word against every other word. It’s called quadratic complexity. As your prompt gets longer, the compute cost explodes. The alternative is the old-school RNN (Recurrent Neural Network). RNNs are incredibly cheap and fast, but they have a fixed memory size. If you give them a long document, they get amnesia. Until today. Google researchers published Memory Caching: RNNs with Growing Memory. And it fixes the biggest bottleneck in AI. Instead of an RNN having a fixed, rigid memory that constantly overwrites itself, Google gave it a "save" button. The technique allows the RNN to cache checkpoints of its hidden states as it reads. The memory capacity of the RNN can now dynamically grow as the sequence gets longer. They built four different variants, including sparse selective mechanisms where the AI actively chooses exactly which checkpoints matter most. The results rewrite the rules of efficiency. On long-context understanding and recall-intensive tasks, these new Memory-Cached RNNs closed the gap with Transformers. They achieved competitive accuracy without the explosive, quadratic compute cost. It perfectly bridges the gap between the cheap efficiency of an RNN and the massive capability of a Transformer. We have spent billions scaling Transformers because we thought they were the only way an AI could remember a long conversation. But Google just proved we don't need to process the whole history every single time. We just needed a smarter cache.

@ComputerrBooks 2

Which work style is best?

@zerohedge Indeed it is. I unpack it in this post: 👇 x.com/i/status/20632…

Payment giants Stripe, Visa, Mastercard said to be among backers of soon-to-debut stablecoin platform: BBG

@SimonDesue Indeed it is. I unpack it in this post: 👇 x.com/i/status/20632…

Visa and Mastercard launching their own stablecoin platform is a bigger deal than most people realize For years crypto has been trying to connect itself to traditional payments Now the payment giants are building the connection themselves Feels like stablecoins are slowly becoming part of everyday financial infrastructure rather than just a crypto product

@CZoom42637 Turbo Pascal

What was your first language, that you printed 'hello world '

Stablecoins are not just a crypto story anymore. Stripe, $V, $MA and $COIN are positioning around something bigger: the interface of money. The financial system is not adopting crypto ideology. It is extracting crypto utility. New piece 👇 open.substack.com/pub/passerlabs…

@SJosephBurns It is rarely a single cause. I think it's a combination of catalysts meeting crowded positioning. My notes below 👇 x.com/i/status/20631…

What do you think triggered the big stock market selloff Friday? 🔽

@0ldskultxo Gonna give a try to AICTX then. Does it run with Codex CLI?

@PasserLabs More the second one: not persistent memory as “remember everything”, but keeping the operational thread alive: decisions, constraints, failed attempts, validation, current state and next action, so the next session continues from what actually happened.

Maybe Coding Agents Don't Need a Bigger Memory. Maybe They Need Continuity. open.substack.com/pub/oldskultxo…

🐦 Sparrow's Weekly read in one minute: 📈 Dow and S&P reached fresh highs. 📉 Friday changed the mood, with the S&P -2.6% and Nasdaq -4.2%. Over $1T erased from semiconductor valuations. Why? Iran remains unresolved. Strong jobs data pushed rate-cut expectations further out. Profit-taking across AI and tech. My read: This looks more like reassessment than panic. The market is taking a breath after a strong run. Observe. Adapt. Move

@clashreport One of the recurring themes in history is the belief that population transfers can solve political problems. They usually create new ones.

Albanian Prime Minister Edi Rama says claims of a secret deal involving Netanyahu and Jared Kushner to relocate Palestinians to Albania have been “weaponized by the enemies of Albania and of Israel.” Rama says the protests are part of what he describes as a “hybrid war” against Albania.

Ups 🫥

@NoahKingJr Nuclear fusion. That would be my bet. Maybe too much?

What’s coming after Quantum Computing?

@TechByTaraa Turbo Pascal

which programming language you learned once but never touched again?

🌡️Pre-Market thermometer, June 5 2026.. Brent holding steady near $94-95. VIX still low and quiet. Premarket mixed. Nasdaq under pressure, Dow soft, S&P near flat. We’re still living inside the “infinite truce” halo with relief being digested. My personal read is: seems we’re on the same line as yesterday. Nothing more to add. Good moment to rebalance your portfolios. No strong bias. Let’s see what the sparrow gets right today. Have a great weekend!

True, you may have a point. 😄 But I think he's right about the broader shift. The good news is that context runtime tools are getting better fast. Solutions like AICTX help manage Codex and Claude sessions much more efficiently, reducing context overhead and token consumption. Worth exploring if you are developing, I personally recommend it to other context engineers. github.com/oldskultxo/aic…

@codingvic Yo creo que nos estamos tragando puro marketing para consumo de tokens, el ha estado en tantas entrevistas ultimamente que dudo que se aparezca en la oficina mas.

Creador de Claude Code: "Ya no le doy prompts a Claude. Escribo bucles - y los bucles hacen el trabajo. Mi trabajo es escribir bucles." en 30 minutos Boris revela su configuración diaria real de Claude Code. Claude Code + bucles + flujo de trabajo dinámico Vale más que un curso de vibe-coding de $500