DeepStack is a general-purpose algorithm - EquiLibre

DeepStack is a general-purpose algorithm

Češi postavili AI, která točí miliardy na burze. Dokud vydělává, neřešíme, jak se rozhoduje, říkají

DeepStack: Expert-level artificial intelligence in heads-up

no-limit poker  -  science.org 

Matej Moravčík,1,2* Martin Schmid ...  

University of Alberta - Ualberta.ca

EquiLibre Technologies — The Next Generation of Algorithmic ...

CTO Rudolf Kadlec a CSO Matej Moravcik

Naučili AI vítězit v pokeru, teď mění (právě asi) obchodování na burze. Hodnota české firmy přesáhla 10 miliard

 

EquiLibre raises undisclosed Series A at $500M valuation, led by Creandum

 

TechCrunch.com 

The DeepMind trio who built a poker AI are now making money for quant hedge funds

https://en.wikipedia.org/wiki/DeepStack  -  Texas hold 'em 

https://arcprize.org/tasks

Jak nám zapnutí dvou nastavení ztrojnásobilo skóre v benchmarku ARC-AGI-3

Prádelna 2030

Jak by mohla možná vypadat něčí prádelna 2030


Takto divně to nakreslil ChatGPT 

 

 

Samsung - Stay Connect

LG - Keep Remote Start ON

Gorenje - Remote Start (24 h)

Bosch - Remote Start (Home Connect)

Haier - Always-on Remote Control (hOn) 

AEG - Remote Start / Remote Start Mode

Siemens - Remote Start / Remote Control (Home Connect)

Hoover - Remote Control (hOn);
přesný ekvivalent „Stay Connect po dokončení“ nemá výrobce jasně pojmenovaný.

 Léto 2026  Unitree 

 

2× tříprstá ruka Dex3-1 (tactile verze), silově řízená (force-controlled)

7 DoF na každou ruku (palec 3, ukazovák 2, prostředník 2)

Každá ruka má 7 DoF a 33 taktilních senzorů, což umožňuje přesnější řízení úchopu a práci s objekty.

 

Německo Čína USA

Mýtus o byrokracii a drahých energiích. Pravý důvod německého úpadku leží v Pekingu


Podle analýzy agentury Bloomberg z konce roku 2024 lze zhruba čtyřicet procent výpadku německého HDP vysvětlit ztrátou exportních trhů, dalších čtyřicet procent vyššími cenami energií a zbylých dvacet procent slabou domácí poptávkou, byrokracií a dalšími faktory. Německá politika tak svým důrazem na byrokracii obrací pravidlo 80:20 naruby.

 

LEAKED SYSTEM PROMPTS FOR
CHATGPT, CLAUDE, GEMINI, GROK, PERPLEXITY, CURSOR, LOVABLE, REPLIT, AND MORE! - AI SYSTEMS TRANSPARENCY FOR ALL!


Jak vypadá systémový prompt v AI Claude.
Tři tisíce slov, které AI přidá ke každému novému dotazu



https://github.com/simonw/research/commits/main/extract-system-prompts





Trumpovo varování se naplňuje: V USA vzniká nová „rudá garda“



Ekonomka Radka Johnová o kolonii Česko: Tato vláda je pořád nejlepší možnost


Evropský svaz? Nekrmte tu bestii!

nomark.project

Emos ventilátor pod radiátor

Do Foods Really Taste Better the Next Day?
A Blinded Sensory Study of Fresh and Day-Old Foods  - PrePrints.org

https://eracompute.com

Coding Factory runs this assembly line

Factory runs this assembly line

This is my personal software factory.

It turns ideas into working products while I sleep. No babysitting coding agents with prompts all day. 

Cursor and Claude Code made writing code way easier. The harder problem is building a system that can context engineer and manage itself.

My factory starts with a Skill called `/factory`. It's the foreman that remembers where the project stands and sends in the right worker for the job. 

Factory runs this assembly line:

1. `/factory-plan` - the interviewer

Reads the existing codebase (if there is one), extracts missing context from me via interview, and writes a product brief. 

2. `/factory-plan` - the planner 

The same skill turns the approved brief into small, testable features and development tasks.

3. `/factory-tests` - the professor 

Before anyone writes code, every task gets an exam. This skill defines the success criteria for each task, and how the coding agent can prove to itself that what it built works or needs iteration. 

4. `/factory-explain` - the presenter

Explains the plan to me like I'm 10, with visual metaphor and mermaid charts. Now that coding agents can write more code, faster than any human ever could, the new bottleneck is human understanding of the code. This skill solves that. 

5. `/factory-handoff` - from CTO to SWE 

This packages the brief, plan, tests, safety rails, and stop conditions into one work order. Factory uses the best models for the planning in the previous steps above, then hands the work order to a lower token usage model like Grok 4.5 for execution.

6. Cursor or Claude Code `/loop` - the coffee 

The night shift picks one task, builds it, takes its exam, records what happened, iterates if needed, then moves onto the next task. If it gets stuck, circuit breakers stop it from confidently digging a deeper hole while I sleep.

7. `/factory-review` - the teacher grades the homework

The student doesn't grade it's own homework. A fresh agent that never met the builder tries to break the result.  The reviewer rereads the original plan, reruns tests, and finds anything that's broken. 

8. `/auto-loom-proof` - shows the evidence 

Uses browser use and screen records itself performing the tests and adds an 11labs voiceover explaining what's being proven. It sends me the narrated demo video. 

9. `/factory-explain` - the code 

The factory updates a plain-language owner's manual explaining what actually got built. I understand my own codebase, so I can make decisions without becoming the bottleneck or outsourcing my thinking to AI. 

NOTE ON BUILDING AI 

Cursor and Claude Code have made writing code dramatically easier. But getting AI to work reliably and at scale for you can't be fully automated.

LLM-as-judge helps, but a judge needs a rubric, examples, and input from someone with subject matter expertise. You still need a human reviewing the work and teaching the system how to perform better.

You can check out my factory on github in the post below.

source: 
https://x.com/mfishbein/status/2081031938228232360




Permanent memory for AI agents. A 426-token prompt, a script, plug and play.
https://github.com/VictorTaelin/OptMem





Graph Engineering Clearly Explained 

from prompt → context → harness → loop → graph engineering:

https://x.com/akshay_pachaar/status/2081089131808243999




https://x.com/toddcohen

https://github.com/squidbay/factory


https://github.com/rennf93/roboco

A universal CLI client for MCP. mcpc supports persistent sessions, stdio/HTTP, OAuth 2.1, tasks, JSON output for code mode, proxy for AI sandboxes, x402, and more
https://github.com/apify/mcpc


Viral Marketing

https://x.com/mfishbein/status/2079991049418940740








3D building editor - Pascal Editor
https://x.com/TechByNagma/status/2081253476374561198

https://github.com/pascalorg/editor

A hive mind communication platform - buzz
https://github.com/block/buzz

Nástroje pro budování komunit - Jaké používáte vy? Jaké s nimi máte zkušenosti?
https://romanripa.substack.com/p/tip-nastroje-pro-budovani-komunit


Maps
Mapy
Mapa

GeoLibre v2.3.0 is here!
https://x.com/giswqs/status/2081066595527348507

https://github.com/opengeos/GeoLibre

Nový život pro šuplíkové mobily

Nový život pro šuplíkové telefony. Webová aplikace ScreenWall je promění v meteostanici či fotorámeček




Webová platforma ScreenWall

ScreenWall | Reuse old phones as synced displays

https://www.reddit.com/r/SideProject/comments/1v66kqo/screenwall_turn_old_phones_into_synced_widgets/

Andrej Karpathy Bilevel autoresearch - Loop Engineering

Loop Engineering: The Karpathy Method - and the workflow that just made it 5x better

Andrej Karpathy autoresearch

https://github.com/karpathy/autoresearch

Bilevel Autoresearch: Meta-Autoresearching Itself

Who knew early singularity could be this fun? :)

I just confirmed that the improvements autoresearch found over the last 2 days of (~650) experiments on depth 12 model transfer well to depth 24 so nanochat is about to get a new leaderboard entry for “time to GPT-2” too. Works 🤷‍♂️

https://x.com/karpathy/status/2030777122223173639

How I Built a Skill That Makes All My Other Skills Better (Using Karpathy's Autoresearch)

I had 20+ skills running my newsletter. The Karpathy Loop showed me most of them were operating at half their potential.

I Turned Andrej Karpathy’s Autoresearch Into a Universal Skill

I generalized Karpathy's autoresearch into a skill for Claude Code. Works on any codebase, not just ML.
 

TheGreenCedar / codex-autoresearch

A codex plugin for running optimization loops inside a codebase. It is useful when you have a measurable target and many possible changes to try: test runtime, build speed, bundle size, model loss, Lighthouse scores, memory use, query latency, or any other metric you can print from a script.

https://github.com/TheGreenCedar/codex-autoresearch


leo-lilinxiao / codex-autoresearch

Codex Autoresearch Skill — A self-directed iterative system for Codex that continuously cycles through: modify, verify, retain or discard, and repeat indefinitely. Inspired by Karpathy’s autoresearch concept.

https://github.com/leo-lilinxiao/codex-autoresearch  

ResearcherSkill 

Cursor/Claude/Codex současně: ResearcherSkill

https://github.com/krzysztofdudek/ResearcherSkill

Desítky paralelních experimentů: evo.

https://github.com/evo-hq/evo




Správné zapojení do workflow

Záměr a akceptační kritéria

        ↓

běžný agent vytvoří funkci

        ↓

autoresearch optimalizuje pouze měřitelnou vlastnost

        ↓

testy a guardy odmítnou regrese

        ↓

člověk zkontroluje nejlepší diff

Jak nastavit měřitelné metriky

Cíl: Co přesně zlepšujeme?

Baseline: Aktuální hodnota.

Primární metrika: Jedno číslo a směr.

Cíl: Hodnota, které chceme dosáhnout.

Benchmark: Jeden opakovatelný příkaz.

Guardy: Co se nesmí zhoršit.

Scope: Které soubory lze měnit.

Rozpočet: Počet pokusů / čas / náklady.

Stop: Cíl dosažen nebo N pokusů bez zlepšení.

Např:

Cíl: Zrychlit import CSV.

Baseline: p95 = 4,8 s.

Metrika: p95 v ms, nižší je lepší.

Cíl: ≤ 3,0 s.

Benchmark: npm run benchmark:csv

Guardy: 100 % testů; shoda výstupu; RAM ≤ 600 MB.

Scope: src/import/**.

Rozpočet: 12 pokusů.

Stop: 4 pokusy bez statisticky významného zlepšení.

Nejdůležitější pravidlo:
jedna optimalizovaná metrika + několik nepřekročitelných guardů. Jinak agent metriku „zlepší“ například vypnutím části práce nebo snížením správnosti. Evo na toto riziko výslovně upozorňuje.





Hub → Spokes → Replies → Insights → Better Hubs

COS OS

I Connected ChatGPT to Typefully and Built an X Content Operating System


Turn ChatGPT into a personal operating system, not a toy. Here’s how I structured it



https://www.generals.bot

https://EquiLibre.ai





https://x.com/pesvklobouku/status/2075228900544655609

Nemluví a je kvadriplegik leukodystrofie Tubb4a - H-ABC

Můj bratr nemluví a je kvadriplegik v důsledku onemocnění zvaného leukodystrofie související s genem Tubb4a, známého také jako H-ABC:
 
My brother is nonspeaking and quadriplegic due to a condition called Tubb4a-related Leukodystrophy also known as H-ABC:



Giving my brother independence again

Co se stalo s Benem?

Chodíš na fyzioterapii?

Jsou ty dlahy na ruce vůbec pohodlné?

Měl něco podobného i Stephen Hawking?

Funguje na úrovni třicetiletého člověka?

Jak Ben jí?

Dá se Benův stav vyléčit?

Bolí ho to?

Jak předcházíte proleženinám?

Jak Ben komunikuje?

Mluví Ben s Elouise?

Opravdu to Ben píše na klávesnici?

Uvažovali jste o Neuralinku?

Je Ben opravdu tak šťastný?

Jak Ben chodí na toaletu?

Proč nosí Ben obličejový štít?

Pořídil si Ben psa?

Jakou hudbu má Ben rád?

What happened to Ben?

Are you doing physical therapy?

Are those arm splints comfortable at all?

Is this what Stephen Hawking had?

Is he operating at the capacity of a 30-year-old?

How does Ben eat?

Can Ben's condition be cured?

Does his condition hurt?

How do you prevent pressure sores?

How does Ben communicate?

Does Ben talk to Elouise?

Is Ben really typing that?

Have you considered Neuralink?

Is Ben really this happy?

How does Ben use the bathroom?

Why does Ben wear a face shield?

Did Ben get a dog?

What music does Ben like?

Revealing the Answers | Life With a Rare Disease
NARBE House

Your Questions Answered
This video goes over your most frequently asked questions about Ben.

https://www.narbehouse.com  

Free games built for switch users

https://www.switchedgames.org 

NARBE Foundation invests in projects that prove people with complex needs deserve autonomy, connection, and choice. We support the builders and share what works

https://www.narbefoundation.org