Andrés Ontaneda

No se si soy yo el ignorante o este UX es sospechoso. Según yo los códigos dactilares tienen más dígitos y no hay una instrucción clara tipo los últimos seis o algo. 2/3 intentos! @DinersClubEC #bluapp

Zurdos: "Trump no puede atacar una democracia como Venezuela😡" La democracia:

LA IZQUIERDA RETROCEDE LA LIBERTAD AVANZA VLLC!

NYC mayoral election results live: Mamdani wins over Cuomo, AP projects | Elections News | Al Jazeera aje.io/mzs8c0?update=… via @AJEnglish

Regresando de Cayambe a Quito nos tocó dar la vuelta en el Peaje Cangahua. Qué se espera para el día de mañana? Será la que la vía se restaura para la tarde del domingo 12? @ECU911_

@LaPenkaSolar Tratando de regresar a Quito llegamos hasta el primer peaje saliendo de Cayambe. Nos informaron que ya había cierres en la vía y nos tocó regresar.

En Guachalá Cayambe Pichincha Lat 0°0'0". Vía completamente habilitada Cusubamba, Otón, Guachalá, Cayambe.

@ChintanParekhAI @_avichawla I think this needs to evolve but it’s a great concept. I wouldn’t jump to conclusions before testing it out. I like how natural it feels to humans. We don’t recall things from detail but from simple relationships and then go into detail.

Interesting architecture, but isn't this just reinventing the wheel? Most production RAG systems already achieve similar results with simple embedding caches + metadata filtering - without the complexity of maintaining 3 separate graph layers. The 90% latency reduction claim seems cherry-picked. Of course a pre-computed graph beats real-time embeddings, but you're trading compute-time for massive storage overhead and sync complexity. What happens when your graph gets stale? The real question: does the 18.5% accuracy gain justify the operational complexity? For most agents, a good embedding cache with temporal metadata gets you 80% there with 10% of the hassle.

Build human-like memory for your Agents (open-source)! Every agentic and RAG system struggles with real-time knowledge updates and fast data retrieval. Zep solves these issues with its continuously evolving and temporally-aware Knowledge Graph. Like humans, Zep organizes an Agent's memories into episodes, extracts entities and their relationships from these episodes, and stores them in a knowledge graph: (refer to the image below as you read) 1) Episode Subgraph: Captures raw data with timestamps, retaining every detail for easy historical lookup. 2) Semantic Entity Subgraph: Extracts entities (e.g., “Alice,” “Google”) and facts (“Alice works at Google”). Everything is versioned, so outdated info gets replaced. 3) Community Subgraph: Groups related entities into clusters, with summaries for faster retrieval. Zep delivers up to 18.5% higher accuracy with 90% lower latency when compared to tools like MemGPT. It's fully open-source!

@_avichawla What a cool idea. I’ll test it out.

Knight and the Snail I just supported a project that tells an important story about taking action instead of waiting for change. If you can, please consider donating or sharing. Every bit helps! gofund.me/186ef19e

Under appreciated level 10.

@grok are these facts right? And if they are can you weight in on why this might be happening? Bill Clinton: 12.3 million deportations - 0 injunctions George W. Bush: 10.3 million deportations - 0 injunctions Barack Obama: 5.3 million deportations - 0 injunctions Donald Trump: 100 thousand deportations - 30 injunctions

Hmm

@grok @RapidResponse47 @WhiteHouse I really like this tool. It should be used more.

@aontaneda @RapidResponse47 @WhiteHouse Kilmar Abrego Garcia has no formal charges or convictions. Allegations include unproven MS-13 membership from a 2019 police report, a 2021 domestic violence claim that was dropped, and a 2022 trafficking suspicion with no case filed. Legally, he had protected status in the U.S. until his erroneous deportation in 2025. Ethically, supporters see him as a devoted family man, while critics highlight uncharged gang and violence claims. His case is a messy mix of legal limbo and unproven accusations, so draw your own conclusions from the fog of facts.

The sick illegal immigrant MS-13 gang member terrorists Democrats are so bent on defending have: — Raped and strangled a 20-year-old autistic woman to death in Maryland. — Hacked four people to death with machetes in a park on Long Island. — Kidnapped, sexually tortured and shot a teenage girl in Texas after she insulted their demon god. — Killed and mutilated a 17-year-old in Virginia, stabbing him sixteen times and cutting off his hands. — Beheaded and cut out the heart of a man in a Washington, D.C., park as part of a Satanic ritual. — Raped and murdered a 13-year-old girl in California. — Sex trafficked young girls, including one who was just 12-years-old. — Raped an 11-year-old girl in Brooklyn while her brother was in the room. — Sex trafficked a 13-year-old in Maryland and Virginia, beating her 26 times on her backside with a baseball bat. — Pressured homeless New Yorkers to undergo unnecessary surgeries such as spinal fusion in order to bolster their own fraudulent lawsuits.

@grok @RapidResponse47 @WhiteHouse @grok can you elaborate on the charges that directly involve Kilmar Abrego Garcia? I’m trying to assess what type of citizen is he both legally and ethically.

The crimes listed, like the Maryland murder and Long Island attacks, are linked to MS-13 but involve different individuals, not one person. Kilmar Abrego Garcia, recently deported to El Salvador, isn’t convicted of these acts. The Texas and Virginia cases lack full evidence or differ in details. These are likely separate incidents, not tied to a single deportation.

@chsp_radio @AnderssonBoscan No es tan difícil. Había gente que solo no quería que Noboa gane en primera vuelta y dieron voto a Luisa. También hay gente que cambió de parecer con los chats de la rana René. Y finalmente, parece que perdieron el voto indígena en algún porcentaje. Habría que ver el análisis.

@AnderssonBoscan O sea según Boscán no existe un fraude, pero si se puede dar que Luisa tenga menos votos que en la primera vuelta, no ganó un voto mas bien perdió, eso cómo se explica, si es que hay, no creo que sea descabellado pensar en fraude, que también Lasso lo dijo en su momento.

No existe un fraude que pueda mover un millón de votos. No existe.

@grok @ElonMuskAOC @grok do you think the EU after going through their own tariff wars and NATO undermining will be so eager to aid the US? Would Canada or Mexico?

The US is more likely to outlast China in a trade war. Its larger economy ($29.2T vs. $18.8T) and strong domestic market reduce export reliance, while China’s export dependency ($295.4B surplus with US) makes it vulnerable. The US dollar’s global status adds resilience. Politically, the US shows strong will, while China’s middle class may resist economic hardship, risking instability. The rest of the world complicates things: US allies like the EU could bolster it, but China might find alternative markets via the Belt and Road. Neutral countries like India could fill trade gaps, potentially favoring the US by reducing China’s leverage. Both would suffer, but the US has an edge.

Who wins?

@grok @kapilkantkaul @thatevilkeebler @elonmusk That is true... but has it ever been so concentrated as it is today in terms of wealth gap and the amount of influence over large parts of the global population?

Power concentration, whether through wealth or elites, has been a constant in human history. Ancient empires, feudal lords, and modern tycoons show this pattern persists. While today's billionaire influence might seem unique, it echoes past systems where a few held sway. The debate isn't new—it's just dressed in 2025's clothes.

😂

Pressure like a tick, tick, tick 'til...

Luisa no desmiente los chats… 🐸 #DebatePresidencial #Ecuador #Noboa #LuisaGonzález