Angel Conde

Senior Solutions Architect at AWS. Ph.D. #IA #IoT #BigData #Cloud and any kind of distributed system. Views are my own.

[Luis Garicano 🇪��🇺🇦 @lugaricano]

Qué país irreformable.
Qué tristeza.

[Angel Conde @Neuw84]

@devruso Y de donde sacas cpu y memoria para que esos modelos hagan algo? ( agentes). Ahora mismo esa parte no esta muy clara en local

[Aitor Murguzur @murggu]

Introducing @muredata, an AI-native data agency.

We build intelligent data systems for companies and the platform that powers it with three pillars: ecosystem reasoning, data infrastructure for agents, and agency runtime.

Feeds is our first beta release. Let’s build, together!

[sarah guo @saranormous]

Caught up with @karpathy for a new @NoPriorsPod: on the phase shift in engineering, AI psychosis, claws, AutoResearch, the opportunity for a SETI-at-Home like movement in AI, the model landscape, and second order effects

02:55 - What Capability Limits Remain?
06:15 - What Mastery of Coding Agents Looks Like
11:16 - Second Order Effects of Coding Agents
15:51 - Why AutoResearch
22:45 - Relevant Skills in the AI Era
28:25 - Model Speciation
32:30 - Collaboration Surfaces for Humans and AI
37:28 - Analysis of Jobs Market Data
48:25 - Open vs. Closed Source Models
53:51 - Autonomous Robotics and Atoms
1:00:59 - MicroGPT and Agentic Education
1:05:40 - End Thoughts

[Andrej Karpathy @karpathy]

It is hard to communicate how much programming has changed due to AI in the last 2 months: not gradually and over time in the "progress as usual" way, but specifically this last December. There are a number of asterisks but imo coding agents basically didn’t work before December and basically work since - the models have significantly higher quality, long-term coherence and tenacity and they can power through large and long tasks, well past enough that it is extremely disruptive to the default programming workflow.

Just to give an example, over the weekend I was building a local video analysis dashboard for the cameras of my home so I wrote: “Here is the local IP and username/password of my DGX Spark. Log in, set up ssh keys, set up vLLM, download and bench Qwen3-VL, set up a server endpoint to inference videos, a basic web ui dashboard, test everything, set it up with systemd, record memory notes for yourself and write up a markdown report for me”. The agent went off for ~30 minutes, ran into multiple issues, researched solutions online, resolved them one by one, wrote the code, tested it, debugged it, set up the services, and came back with the report and it was just done. I didn’t touch anything. All of this could easily have been a weekend project just 3 months ago but today it’s something you kick off and forget about for 30 minutes.

As a result, programming is becoming unrecognizable. You’re not typing computer code into an editor like the way things were since computers were invented, that era is over. You're spinning up AI agents, giving them tasks *in English* and managing and reviewing their work in parallel. The biggest prize is in figuring out how you can keep ascending the layers of abstraction to set up long-running orchestrator Claws with all of the right tools, memory and instructions that productively manage multiple parallel Code instances for you. The leverage achievable via top tier "agentic engineering" feels very high right now.

It’s not perfect, it needs high-level direction, judgement, taste, oversight, iteration and hints and ideas. It works a lot better in some scenarios than others (e.g. especially for tasks that are well-specified and where you can verify/test functionality). The key is to build intuition to decompose the task just right to hand off the parts that work and help out around the edges. But imo, this is nowhere near "business as usual" time in software.

[Andy Coenen @_coenen]

I wanted to share something I built over the last few weeks: https://t.co/QRqMK9CpTR is a massive isometric pixel art map of NYC, built with nano banana and coding agents.

I didn't write a single line of code.

[Angel Conde @Neuw84]

Bonita visualización del tráfico en Foronda a través de señales ADBS.

https://t.co/OOijRsAoa5

@ForondaAirport @AeroclubVitoria

[Angel Conde @Neuw84]

@javisantana Go ahead. No super experto eso si!

[Tuta @TutaPrivacy]

🚨Our governments are about to decide whether 450M Europeans deserve privacy - or not.

Help ensure your country says NO to Chat Control: Call you local representatives!

Privacy is not negotiable. Speak up now. ✊ #privacy

👉 More on how to stop Chat Control: https://t.co/JLei3q4sLi

[Angel Conde @Neuw84]

@JaVidalPe Entonces sí, no te merece la pena. Parece que tienes limitación de computo y lo que estabas haciendo es lo correcto ( por si acaso sugería la prueba!)

[Angel Conde @Neuw84]

@JaVidalPe Entiendo que por temas de io? No si es así lo más fácil es cambiar a las io optimized (optimizacion de queries aparte)

https://t.co/jPdcXHjQm3

[Nick Erickson @innixma]

AutoGluon 1.0 is live!! Shatters SOTA, wins 75% vs prior release, 63% win-rate vs best-in-hindsight combination of other methods.
To our knowledge, this is the biggest leap forward in tabular ML in the past 4 years.

See how we did it: https://t.co/CPNTdcNH99

#AutoML #AutoGluon

[Confluent @confluentinc]

📣🐦 Don't miss out on our early bird pricing for #KafkaSummit Bangalore! Secure your spot by March 15 and save. Join industry leaders, experts, and fellow Kafka enthusiasts at the must-attend data streaming event. Register now -> https://t.co/95XBfK6toM

[Angel Conde @Neuw84]

@javisantana Tanto Amazon como Microsoft le estaban dando bastante apoyo (con sus respectivos providers), a ver qué pasará ahora. A mi siempre me ha parecido complejo el tema (e.g. elastic y el hacer negocio sobre Apache Lucene, sobre todo cuando empezaron)

[Angel Conde @Neuw84]

@murggu What a week of AI announcements!

[Angel Conde @Neuw84]

@seostratega .

[Angel Conde @Neuw84]

I just published Securing Amazon OpenSearch Service Dashboards with Amazon Cognito https://t.co/ss4Fvzxx9h

[Angel Conde @Neuw84]

@IKERLANofficial hablando sobre AutoML usando #sagemaker @awscloud_es

[AWS España @awscloud_es]

El 24 de marzo te contamos qué entendemos en AWS por AutoML, damos un repaso a las distintas herramientas disponibles para distintos perfiles de usuarios, así como crear y desplegar modelos de ML de alta calidad usando las herramientas de AWS de AutoML ✅ https://t.co/uGrdZ2JEZO

[Angel Conde @Neuw84]

@rakyll A lot of people just try to do big data with default parameters without knowing how much performance can be gained tuning the jvm!