UN CIENTÍFICO DANÉS PROGRAMÓ A CLAUDE PARA QUE BUSQUE TRABAJO POR ÉL Y LO ACABA DE HACER PÚBLICO

Mandar CVs es uno de los trabajos más absurdos del mundo: copiar, pegar, adaptar, personalizar, repetir. Todo manual, todo lento, todo para que lo lea un algoritmo antes que un humano.

→ Analiza la oferta de trabajo automáticamente
→ Genera un CV personalizado para cada puesto
→ Redacta la carta de presentación adaptada al contexto
→ Todo lo hace Claude por debajo, sin que toques nada
→ Open source, ya en 3.5k stars en GitHub

El tío que debería estar buscando trabajo ha construido la herramienta que lo busca por él.

Aquí te explico cómo funciona 👇(repoo al final del hilo)

For one year, one study was the strongest evidence in the world that ChatGPT helps students learn.

Nearly 500,000 accesses. Over 500 citations. Altmetric score above 1,000. Universities used it to justify putting AI in classrooms. Policymakers pointed to it when asked if generative AI belongs in education.

On April 22, 2026, the journal retracted it. The evidence was flawed. The conclusions were invalid.

The paper was titled "The effect of ChatGPT on students' learning performance, learning perception, and higher-order thinking." Published May 2025 in Humanities and Social Sciences Communications, a journal in the Nature Portfolio. The authors, Jin Wang and Wenxiang Fan of Hangzhou Normal University, synthesized 51 studies and concluded that ChatGPT had a "large positive impact on improving learning performance" and a "moderately positive impact on enhancing learning perception and fostering higher-order thinking."

Their recommendation was direct. ChatGPT should be "actively integrated into different learning modes to enhance student learning."

Schools listened. Edtech companies cited it. The paper ranked in the 99th percentile for online attention among academic articles.

Then two researchers at UiT Norges arktiske universitet, Magnus Ingebrigtsen and Marko Lukic, went through the meta-analysis line by line. They found a detail that should have ended the paper on the day it was submitted.

The most heavily weighted study in the analysis had already been retracted before Wang and Fan published theirs.

Nature let it through peer review anyway.

The Norwegians filed with Springer Nature's research integrity group in April 2026. The article was retracted five days later. The retraction notice reads: "The Editor has decided to retract this paper owing to concerns regarding discrepancies in the meta-analysis. These issues ultimately undermine the confidence the Editor can place in the validity of the analysis and resulting conclusions. The authors have not responded to correspondence regarding this retraction."

Ben Williamson at the University of Edinburgh, one of the first researchers to publicly flag the problem, put it plainly. "It really seemed like a paper that should not have been published in the first place."

Here is the part that should sit with you.

The retraction was quiet. The paper was viral. Williamson said the headline finding will likely keep circulating on social media, stripped of its methodological caveats, long after the correction.

Over 500 papers already cite this study. Those citations are now embedded in other research. Other conclusions. Other policy recommendations. The retraction does not reach backward. The flawed finding is already inside the literature, being cited by papers that do not know it has been pulled.

Meanwhile, the studies showing that AI reduces critical thinking, creates cognitive dependency, and degrades skills when removed have not been retracted. They keep replicating. Across MIT, Stanford, Zurich, and The Lancet.

The strongest evidence that ChatGPT helps students learn was pulled. The evidence that it hurts them still stands.

14 GitHub repos to boost your software engineering career

(bookmark + star them now):

Fable5を効率活用する神手法

Claude Fable 5 + Opus + Sonnet + Codex = 開発チームを丸ごと再現する司令塔システムが海外で話題です。

このAIオーケストレーション術では、Anthropicの最新最上位モデル「Fable 5」を司令塔に据えます。実際の開発作業は、Opus・Sonnet・Codexという3人の専門部下に割り振ります。

一番高いモデルで全部の作業をこなす必要なし。バグ調査から単純作業まで一人で抱え込む必要もなし。判断に迷って時間を溶かす必要もなし。

Fable 5の高い利用コストを節約しながら、複雑な開発をチーム運営に任せられます。

仕組みはこうです👇

scrcpy mirrors and controls Android devices from your desktop over USB or TCP/IP with low latency, no root required, and no app installation on the device.

- 30–120fps mirroring at up to 1920×1080 or above
- Audio forwarding (Android 11+), recording, and virtual display support
- Physical keyboard and mouse simulation via HID
- Works on Linux, Windows, and macOS

Explore it here:
https://www.opensourceprojects.dev/post/scrcpy