Bindu

solving problems from first principles

[王多鱼AI @wdy188]

做 AI 算命、做命理 App 的朋友，这份 57GB 的开源数据可以直接拿去用。

据我们所知，这是目前全网最大的玄学类开源数据集：1950–2026 年每一天 × 12 个时辰 × 男女，67.5 万张紫微命盘全排了一遍，13 个维度一共 877 万段解读。

每一句���语都对照倪海厦《天纪》原话和《紫微斗数全书》原文校正过，引文逐字可查。全部按规则算出来，不经 AI 生成，同一张盘每次结果都一样。

压缩后只有 161MB，按出生年份一年一个文件。CC BY 4.0，可以商用，细则见仓库 README。

搭排盘网站直接查表，训练命理大模型直接当语料。仓库已经 5,400 多颗星，欢迎拿去用。

🔗 https://t.co/CSIjzShojj

[Piyush Sachdeva - ReelDrop #1 Instagram GrowthTool @techie_piyush]

🚨 THE INTERNET BROKE TODAY 🚨

Oct 21, 2026 will go down in history

1. eMudhra : Private key was literally "4". 8 keys forged ANY website for 17 months. 60M+ certs at risk.

2. SBI CONFIRMED. 3,897 customers' mobile, DOB, address + loan data found on dark web.

3. Rajasthan Police 🇮🇳ALLEGED hack. Group claims 70,000 identity records + 59,460 numbers for sale. Unverified.

4. php2shell [9.2] Unauth RCE in PHP. 70% of the web. WordPress, Drupal, Laravel. CISA: Patch or GO OFFLINE.

5. GitLab [10.0/10] PERFECT SCORE. Unauth RCE, no login needed.

6. Node.js [CVE-2026-111497] Brotli decoder RCE + TLS bypass. 10 CVEs patched today. Also blocked eMudhra.

7. Railway BREACHED TODAY. Env variables decrypted. Your API keys & DB passwords exposed. Rotate all.

8. OpenCode BREACHED. OpenAI, Anthropic, Google API keys exposed for Console users. Distillation attack. Bills run up in $$$$.

9. LiteLLM PyPI v1.107.0 was MALWARE today morning. Steals credentials + backdoor.

10. Plex2Shell ACTIVELY EXPLOITED. Hackers installing cryptominers LIVE. CISA KEV, deadline Oct 23.

AI found the bugs. AI is exploiting them. Humans can't patch fast enough.

Patch. Rotate. Revoke. Update.
RT to save someone's server ♻️

#CyberSecurity #DataBreach #India #php2shell

[Gresham @Greshamtrade]

holy sh*t. a 17-year-old computer science student from China runs 300 Grok Bots for less than $0.02 per bot a week

the same swarm would cost $4,500 a week. he cut empty checks, stopped rereading entire chats, and killed retry loops. most bots now wake up only when there’s actual work to do

20 GitHub repos he used to make the setup cheaper:

01 LiteLLM → track spending and set budgets for connected API calls before one runaway workflow eats the whole allowance
https://t.co/jGFr8MccPo

02 RouteLLM → route simpler API requests to a cheaper model and reserve the stronger one for work that needs it
https://t.co/TJnBuQVqdz

03 Semantic Router → match requests to predefined workflows using local embeddings instead of paying a large model to decide every next step
https://t.co/XRURo0i2Xq

04 GPTCache → reuse suitable answers to similar requests instead of generating them again. define matching and freshness rules around the task
https://t.co/k1XwyJvPgs

05 requests-cache → keep fetched responses with expiry rules, then let your scripts reuse valid data before asking the bot to investigate again
https://t.co/TFMJrGj50V

06 Hishel → add HTTP caching and revalidation to your fetching tools, keeping unchanged responses out of repeated download-and-analysis cycles
https://t.co/XFoWwTc33T

07 Crawl4AI → turn web pages into clean Markdown before analysis, reducing the navigation menus and page clutter your bot has to read
https://t.co/HTuqEY4NI3

08 Firecrawl → collect web content as Markdown or structured data so your workflow can select the fields it actually needs
https://t.co/EJKKIAeH7x

09 Chonkie → split long documents into manageable chunks so retrieval can select relevant passages instead of attaching the whole document
https://t.co/RhokoGmbpq

10 FlagEmbedding → rerank retrieved passages before generation, keeping the strongest evidence and dropping irrelevant matches from the prompt
https://t.co/1cRBETSfgP

11 LLMLingua → compress lengthy inputs before the model reads them. preserve the facts, constraints, and qualifications your task depends on
https://t.co/ZLGYBf9FOU

12 RTK → compress common command outputs before they enter the conversation, keeping repetitive logs from swallowing the useful context
https://t.co/cu7IuXOuyx

13 Mem0 → store and retrieve relevant facts across runs so each assignment can start with selected memory instead of the entire conversation
https://t.co/yyWL8KDCwI

14 Graphiti → keep changing facts and relationships in a searchable graph, then retrieve the relevant context instead of replaying the event history
https://t.co/89yr3C13x7

15 Prefect → cache successful task results and reuse them when the configured inputs match, avoiding unnecessary reruns of completed workflow steps
https://t.co/QSw5vpLOOb

16 Celery → put background jobs on a worker queue so code handles waiting and dispatch while model calls happen inside the tasks that need them
https://t.co/ILjsLdssce

17 APScheduler → schedule lightweight checks in Python and batch nonurgent work before handing useful updates to the agent
https://t.co/RC8DDERSNg

18 PyBreaker → pause calls to a repeatedly failing service after a configured threshold, stopping one broken dependency from generating endless attempts
https://t.co/XqS8M1HMAm

19 Portkey Gateway → cache matching API responses behind a shared gateway, letting connected workflows reuse answers across repeated requests
https://t.co/Dg8TdEdj40

20 Phoenix → trace your instrumented AI workflow, inspect repeated calls, and compare changes against output quality before keeping an “optimization”
https://t.co/0nJC2XkoLL

connect the tools your workflow needs through the terminal, MCP, or your own code. route API calls through the gateway or router you choose so its controls actually apply

start with the biggest leak: oversized inputs, repeated research, empty checks, or failed operations that keep restarting

filter before reading → reuse before generating → check for changes → stop failed loops

then measure completed work against total usage, including the tools running around your bot

save this before you give the same waste a bigger budget

the full course below shows how to put the workflow together ↓

[Ruuj @RuOnChain1]

I’VE NEVER SEEN AN OPEN-SOURCE TRADING FRAMEWORK ENGINEERED THIS SERIOUSLY.

a complete multi-agent LLM trading framework

agents analyse, debate and review risk before every decision

here’s how I’d build a strategy factory around agents like these:

[Brendan Jowett @jowettbrendan]

GitHub is a GOLDMINE for AI game developers.

I picked 10 powerful open-source repos that make building 3D games with Claude Code and Codex ridiculously easier.

Save these before your next project:

1. Gamedev Agent Skills

Give Claude Code and Codex reusable game-development skills for engines including Godot, Unity, Unreal, and Three.js.

https://t.co/WudWTsajr7

2. MCP Unity

Connect your coding agent to Unity. Create scenes, modify objects, assign materials, run tests, and read console errors through prompts.

https://t.co/cosCfK56AY

3. Godot MCP

Let your AI assistant create scenes, add nodes, run Godot projects, and inspect debug output while building your game.

https://t.co/qXwHTtguqL

4. TripoSR

Turn a single image into a 3D model. Useful for prototyping props and exploring assets before refining them for your game.

https://t.co/E9dR1g43n2

5. Drei

Ready-made helpers for React Three Fiber: camera controls, environments, skies, shadows, and model loading for browser-based 3D games.

https://t.co/eDpvF3mZUm

6. React Three Rapier

Add physics to React Three Fiber games, including rigid bodies, collision detection, and joints, without writing a physics engine yourself.

https://t.co/gkYgrb3cC5

7. Yuka

Build NPC behavior with steering, pathfinding, perception, and state-driven decisions. Give your characters more to do than stand around.

https://t.co/YCw5RHiFiQ

8. Three.quarks

Build particle effects for Three.js games, from sparks and smoke to magical trails. Includes a visual editor for designing effects.

https://t.co/ycVaphWBB4

9. React Postprocessing

Give React Three Fiber scenes a visual upgrade with bloom, depth of field, noise, and vignette effects.

https://t.co/DFjsI0swlP

10. Colyseus

Build multiplayer features with rooms, matchmaking, reconnections, and automatic game-state synchronization between your server and players.

https://t.co/UlH13ZuIQJ

Which one are you using in your next game?

[liam. @liambraus]

ALGUIEN EN GITHUB HA SACADO UNA API DE WHATSAPP GRATIS

ahora puedes ejecutar una api de WhatsApp en tu propio servidor sin pagar a Meta.

se llama OpenWA. 12.8K estrellas.

la api oficial de WhatsApp Business se paga y te ata a su plataforma. OpenWA te da automatización por api, autoalojada y gratis.

→ envía y recibe mensajes con una api rest
→ ejecuta varias sesiones de WhatsApp a la vez
→ gestiona sesiones y webhooks con un panel web
→ conéctalo a n8n, Chatwoot y agentes de ia con MCP

clonas el repo, un comando de Docker y el panel está en marcha.

ojo: es no oficial, usa un número dedicado.

100% gratis. código abierto.

os dejo el repo abajo.

[Supabase @supabase]

Supabase is now fully available and open source thanks to Claude

[Bill Gross @Bill_Gross]

Please take a look at the idea database, and if you need funding, reach out to us as well. We want to help get more startups created to solve important problems..

[Alex Nguyen @alexcooldev]

Building casual games will be the next big trend.

No AI costs. No API costs. Just store everything directly in users’ iCloud.

And there are so many ways to monetize: rewarded ads, in-game purchases, remove-ads packages, and more.

But I still have no idea how to market my game lol.

I’ll try organic first before running paid ads.

[Sam Pasupalak @spisallyouneed]

Introducing SKY-001, the first Engineering World Model.
On CAD and BIM tasks, it beats GPT-6 Astra and every frontier model we tested:

-> 51-125x faster
- > 150-210x cheaper
-> Higher IOU than all frontier models

SKY-001 lets you dream a physical object before you build it.

18 months of work. Built by the team out of Maluuba's deep learning lab (acq. by @MSRResearch).

The first ‘AI General Engineer’ will be the backbone for Autonomous Manufacturing.

[Yacine Mahdid @yacinelearning]

if you are feeling hella lost with what’s going on in ML well buddy I have a 1h08min interview with the legendary @vboykis that should set you straight

in it we got:
- good dose of pragmatism
- what you should still learn
- focusing
- how to not lose your braincells to agents

[Bindu @findbindu]

We are not the pinnacle of evolution. The forest reminds us that intelligence, communication, and cooperation exist far beyond the human mind. Nature is far more complex than we give it credit for.

#Nature #Forest #PlantIntelligence #Neuroscience #Evolution #Consciousness

[0xMarioNawfal - The Roundtable Show @RoundtableSpace]

Epic Games released an official plugin that lets Claude Code control Unreal Engine directly.

It connects Claude Code to the Unreal Editor and exposes hundreds of editor tools across 30+ toolsets:

→ Actors, Blueprints and materials
→ Niagara effects, Control Rigs and Sequencer
→ State Trees, widgets and the Gameplay Ability System
→ Automation testing

Install it from Anthropic's official plugin marketplace:
/plugin install unreal-engine-skills-for-claude-code@claude-plugins-official

In Unreal, enable the ModelContextProtocol and AllToolsets plugins, then start the MCP server with the editor open.

Works on Mac, Linux, and Windows through Git Bash or WSL.

https://t.co/LYM4whQ8Ev

[Janis Hesse @HesseJanis]

Very excited to share the most important discovery of my career yet: your brain switches between a conscious and an unconscious mode ~15 times every second. Consciousness may depend not on where in the brain activity occurs but when.
Find out more below.
https://t.co/23dyMGnaLX

[Nirmalya Kajuri @Kaju_Nut]

"Our initial evaluation suggested that GPT-6-Astra's score was in the late eighties followed closely by Opus 5.5. But, as we started to investigate the mistakes that the models might have made, the models, especially Astra, insisted that they were right and that it was the expert-supplied answers that were erroneous. When we went back to the experts it turned out, in multiple instances, that that the experts found some subtle errors buried deep inside their calculations. As they revised their answers, Astra's score started to inch up and it finally ended up with 98% (49/50)."

[Tatsunori Hashimoto @tatsu_hashimoto]

Deep learning is often taught as a sanitized collection of facts and math. But work at the frontier of AI is messy and empirical.

Our new class, Deep Learning Alchemy, aims to teach the art and science of deep learning in the hands-on spirit of Language Modeling From Scratch.

Students build intuition by running experiments on small LMs, then test it through prediction quizzes: given an unseen code diff, can you predict what happens?

https://t.co/0H82T1yyam

My co-instructor @kothasuhas and I think there's significant demand for classes that treat students as serious learners who want to be assessed for what they’ve learned, and so we're trying a new exam-heavy approach to the class (7 exams, 85% of the grade, with the remaining 15% in-person discussion assessments).

We’ve designed the class so students can use AI agents to run their experiments and do their assignment, but they then have to demonstrate that mastery on exams without access to the assignment, and any partial credit on the exam has to come through citations to the assignment, incentivizing them to know what’s in their write-up.

The same assignments and quizzes are public for anyone to try (lecture recordings will be released through @StanfordOnline with some delay). We have some very fun assignments and quizzes covering both classic topics (like NQM, run-to-run stochasticity, and scaling laws) and new and timely topics like muP or hyperball. Made by some real experts like @TristanThrush and @wen_kaiyue

Try some quizzes below!

[Brais Moure @MoureDev]

Si estás buscando crear animaciones para explicar algoritmos o matemáticas, acaban de lanzar este motor 3D en Python de código abierto.

Está pensado especialmente para usarlo junto con agentes de IA, y al parecer es mucho más rápido que Blender.

→ https://t.co/kIbhnb8ajc

[Mathieu @miniapeur]

Fields medallist Hugo Duminil-Copin was deeply shaken by OpenAI’s release of more than 700 AI-generated mathematical manuscripts, including work on 16 questions directly relevant to his research.

���On Tuesday night, I cried, and it woke up my daughter”

“They have annihilated my field of research.”

https://t.co/RQliNgxjS8

[imjustnewatai @imjustnewatai]

Things are about to get alien.
Look at this proof. The answer isn't the crazy part. It's how it got there. The key step connects two ideas that no human thought were connected. Mathematicians reading it call it alien math. It works, it checks out, and it still feels like it came from somewhere else.
This is the beginning of alien math. Proofs that are correct but don't think like we do.
And it won't stop at math. Alien physics is next. New ideas about how reality works that no human would have come up with. Theories that work but feel like they came from another civilization.
After that, alien chemistry. Alien materials. Alien medicine. Alien engineering.
We're not just getting faster science. We're getting a different kind of science.
This is the least alien it will ever be.

[Muthukrishnan Dhandapani @dmuthuk]

I have said this many times before. America is no longer what it used to be.

India trusted Trump and his cronies, Musk included, for far too long before their true colours showed.

Trump would not be President today without Elon Musk’s money and his personal campaign on Twitter.

Musk spent close to $290 million and then claimed the election could not have been won without him. That influence was real.

When the relationship with Trump soured, Musk posted that Trump is in the Epstein files and that this was why they had not been released. What kind of person makes someone President while knowing what those files contain?

Musk is a great technocrat and entrepreneur.
That does not excuse this.

Is “Dear Prime Minister Ambani” how you address the world’s largest democracy and a sovereign nation?

He accuses Ambani of being an oligarch. By that standard, Musk is one too.

Which party we support or oppose does not matter. In a democracy we all have our preferences. When it comes to Indian sovereignty, we are one nation.

America wants India to follow its tantrums. Obey, and do not argue.

It does not like India having an independent view or strategy.

India wants to be treated as an equal.

America wants India as a vassal.

That is why the elites in the American state have turned hostile towards India.

They already regret how far China has risen. They do not want India to reach that point.

In the coming years America will do what it can to harm India’s interests.

We need to face this together as a nation. Very tough times ahead.