Juan D Alfonso

[F.O.L.A @folaoftech]

She has no clue what her son is up to… honestly, I don't get it either.

[Juan D Alfonso @jdalf]

@karthikponna19 Visual basic

[NZ ☄️ @CodeByNZ]

In 2025, this is the minimum requirement for landing a junior developer role:

[Juan D Alfonso @jdalf]

@CodeByNZ Get ready, cause here you have the minimum requirements for 2026.

[Harrison Kinsley @Sentdex]

People often ask how did the Unitree robots get so good all of a sudden.

It wasn't all of a sudden, and it's because they ship their hardware and open source their SDKs. Arguably these robots are nearly useless out of the box, but you have full dev control of them.

Because of that, the hardware has become a very popular R&D platform with an ecosystem around it and the Unitree G1 is undoubtedly an order of magnitude better than it could ever be at this point if Unitree was instead just doing quiet internal dev of both the hardware and software.

Too many hardware companies for really cool products that seek to be community-driven (robots, AR glasses...etc) desire to make a profitable walled garden and their greed just ends up walling out developers and their product gets outpaced by the G1s of the world.

[Massimo @Rainmaker1973]

Scientists have mapped the entire neural network of the fruit fly.

50 million synapses between 139,000 neurons.

[Anthropic @AnthropicAI]

Today, we announced that we plan to expand our use of Google TPUs, securing approximately one million TPUs and more than a gigawatt of capacity in 2026.

[Juan D Alfonso @jdalf]

remote setup for serious code.

[Juan D Alfonso @jdalf]

@midudev

[Juan D Alfonso @jdalf]

En el Louvre usaban Windows XP y contraseñas como “LOUVRE” para los servidores de cámaras. La historia se repite como el robo del siglo 👀
Hora de recordar que aprender ciberseguridad no es opcional 😅Esto se lo dejo a alguien con más alcance, como @freddier y @chemaalonso .

[Juan D Alfonso @jdalf]

En el Louvre usaban Windows XP y contraseñas como “LOUVRE” para los servidores de cámaras. La historia se repite como el robo del siglo 👀
Hora de recordar que aprender ciberseguridad no es opcional 😅Esto se lo dejo a alguien con más alcance, como @freddier y @chemaalonso .

[Juan D Alfonso @jdalf]

@_devJNS This beauty… no ads, no tracking, just pure lag. Just perfect.

[Camilo Andrés García @hyperconectado]

HILO | Fuga de Datos de la empresa de logística Envía: 140 millones de registros de clientes colombianos en riesgo por presunto ciberataque.

[Juan D Alfonso @jdalf]

@hipdead010 Años sin hacer esto y me impresiona que todavía sigan regalando apiKeys como si fueran stickers 😂. Para los nuevos: hacer spiders y regex para rastrear claves es pan comido. Pero bueno, así es trabajar en Tech sin entender el lado Tech 🤷‍♂️

[Juan D Alfonso @jdalf]

En el Louvre usaban Windows XP y contraseñas como “LOUVRE” para los servidores de cámaras. La historia se repite como el robo del siglo 👀
Hora de recordar que aprender ciberseguridad no es opcional 😅Esto se lo dejo a alguien con más alcance, como @freddier y @chemaalonso .

[Juan D Alfonso @jdalf]

@_devJNS Not for seniors.

[Python Developer @PythonDvz]

Top 10 Loss Functions in Machine Learning (and when to use them)

Regression Losses
1️⃣ Mean Bias Error (MBE) – Captures average bias in predictions. Rarely used since positive and negative errors cancel out.

2️⃣ Mean Absolute Error (MAE) – Average absolute difference between predicted and actual values. Treats small and large errors equally since gradient magnitude is constant.

3️⃣ Mean Squared Error (MSE) – Squares errors, making large errors count more. Useful, but sensitive to outliers.

4️⃣ Root Mean Squared Error (RMSE) – Square root of MSE. Keeps loss in the same units as the target variable.

5️⃣ Huber Loss – Hybrid of MAE and MSE. Acts like MSE for small errors and MAE for large ones. Needs a hyperparameter to define the transition point.

6️⃣ Log-Cosh Loss – Smooth, non-parametric alternative to Huber. More stable but a bit more computationally expensive.

Classification Losses

1️⃣ Binary Cross-Entropy (BCE) – Standard for binary classification. Measures mismatch between predicted probabilities and true labels.

2️⃣ Hinge Loss – Based on the margin between points and decision boundary. Penalizes wrong predictions and low-confidence correct ones. Used in training SVMs.

3️⃣ Cross-Entropy Loss – Generalization of BCE for multi-class classification tasks.

4️⃣ KL Divergence – Measures how one probability distribution diverges from another. For classification, minimizing KL is equivalent to minimizing cross-entropy, but it’s widely used in t-SNE and knowledge distillation.

[Weaviate AI Database @weaviate_io]

Your RAG system is probably broken.

Here's how to fix it in 2025.

(𝗔𝗹𝗺𝗼𝘀𝘁!) 𝗘𝘃𝗲𝗿𝘆 𝗔𝗜 𝘁𝗲𝗮𝗺 𝗵𝗮𝘀 𝘁𝗵𝗲 𝘀𝗮𝗺𝗲 𝗽𝗿𝗼𝗯𝗹𝗲𝗺: their RAG retrieves irrelevant chunks, hallucinates answers, and performs worse than expected.

Here’s how to fix it:
𝗦𝘁𝗼𝗽 𝘁𝗵𝗶𝗻𝗸𝗶𝗻𝗴 𝗿𝗲𝘁𝗿𝗶𝗲𝘃𝗮𝗹, 𝘀𝘁𝗮𝗿𝘁 𝘁𝗵𝗶𝗻𝗸𝗶𝗻𝗴 𝗮𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲.

These RAG patterns are dominating production systems:

• Naive RAG - https://t.co/Y9tiWPrxKo

• Agentic RAG - https://t.co/7HiZtHqjie

• Advanced RAG - https://t.co/hTqkDV8cGX

• Multimodal RAG - https://t.co/EPBHqbULub

• Graph RAG - https://t.co/rHgunUumdX

• Retrieve-and-rerank - https://t.co/BDidkwwI8e

𝗧𝗵𝗲 𝗰𝗼𝗺𝗽𝗮𝗻𝗶𝗲𝘀 𝗰𝗿𝘂𝘀𝗵𝗶𝗻𝗴 𝗶𝘁 𝘄𝗶𝘁𝗵 𝗥𝗔𝗚?

They're orchestrating these patterns, not just chunking documents.

Master your architecture with these FREE ebooks:

📌 Advanced RAG Ebook: https://t.co/nmdhvNrt8S
📌 Agentic Architectures Ebook: https://t.co/xE9YjRDjtB

Start here: Pick ONE pattern that solves your biggest retrieval pain point. Don't try to implement everything at once.

[FAR.AI @farairesearch]

Can jailbreaking AI be prevented with signal processing techniques?

@pinyuchenTW shares a unified framework treating AI safety as hypothesis testing. Unlike methods with predefined parameters, safety hypotheses are context-dependent, requiring language-model-as-a-judge
👇

[NO CONTEXT HUMANS @HumansNoContext]

Minimalism took away the glory

[JNS @_devJNS]

actors so good they convinced an entire generation that coding is cool asf.