David Fombella (@davidfombella) header
David Fombella (@davidfombella) profile picture
Oviedo/Madrid
Joined January 2009
Photo by David Fombella (@davidfombella)Photo by David Fombella (@davidfombella)Photo by David Fombella (@davidfombella)Photo by David Fombella (@davidfombella)Photo by David Fombella (@davidfombella)Photo by David Fombella (@davidfombella)Photo by David Fombella (@davidfombella)Photo by David Fombella (@davidfombella)Photo by David Fombella (@davidfombella)Photo by David Fombella (@davidfombella)
David Fombella retweeted
I bet you missed these new things in Power BI 👀 Power BI has shipped a bunch of upgrades to the matrix visual over the last few months, and most people I talk to haven’t noticed them yet. The poster shows all 5 and where to find each one. My favourite is the column width control, along with changing totals without writing any DAX. Both used to annoy me so much. Now I enjoy fixing them. Which one is your favourite? Tell me in the comments 👇 Follow me for more Power BI tips and tricks. And tag a friend who’d love these updates 🤩
2
6
80
2,770
David Fombella retweeted
HIMNO DE ESPAÑA EN #MADRING ❤️💛❤️ Qué orgullo. Qué afortunados somos de tener dos pilotos españoles. Qué suerte tenemos de vivir dos carreras en nuestro país 🥹👏 #EspañaDAZNF1 🇪🇸
231
2,894
291
14,573
1,498,186
David Fombella retweeted
Skills are context; they don't replace tools. When experimenting with agentic Power BI report development last year, the first successful approach was a skill I made. However I abandoned this (and later shared it) because it's unreliable & expensive. Compare a pbir-cli command with generated JSON for one visual. Amplify this across many visuals, pages, reports...
2
1
30
2,292
David Fombella retweeted
A senior Google engineer dropped a 424-page doc on agentic design patterns. 424 pages. Most engineers bookmarked it and never opened it again. I read the whole thing. Here are the 15 patterns that actually matter — explained in plain English, with exactly when to use each one ↓
54
478
8
2,578
507,228
David Fombella retweeted
I've created a benchmark that measures agent performance in basic Power BI report tasks, showing that directly modifying PBIR files is worse, more expensive, and (often) slower than using pbir-cli. The tests evaluate 30 simple tasks like "create a card that shows the total invoice lines, using <tool or skill>" and a pass is when the agent can do it correct across 3/3 replicates. So if it does it 1 or 2 times, it still fails. The results were quite interesting. A plugin with skills and small CLIs for validation didn't meaningfully increase performance much (esp not for Opus 4.8), but ballooned token cost and time. In contrast, the pbir-cli alone was sufficient to increase performance significantly on those tasks, and adding the skill made it faster and take less tokens. "Dumber" models took longer with the CLI because they spent more time exploring or iterating, and seem worse at using skills in general. Obviously together with @maxanatsko I'm an author of the pbir-cli so it's appropriate to be skeptical of my benchmark that favors my own tool, but I did try to make these tests objective and fair. Note this benchmark only evaluates performance as a function of "can it complete the task" not "do the report / visuals look nice". I'm building a design evaluation as well but it takes a lot longer since design is subjective and coming up with good, repeatable tests that aren't handwavy marketing bs (like "make the report from this image") is challenging.
5
8
2
48
4,122
David Fombella retweeted
You have noticed it. ChatGPT feels dumber than it used to. Your prompts that worked six months ago produce worse results now. The writing sounds flatter. The ideas sound safer. The internet itself feels like it is shrinking. Every article reads the same. Every email sounds the same. Every answer sounds like it was written by the same voice. You thought it was you. It is not you. Researchers at Oxford and Cambridge published a paper in Nature proving what is happening. They call it Model Collapse. Here is the mechanism in one sentence. AI trained on AI-generated data gets dumber every generation until it forgets what real human data looked like. The internet is filling with AI-generated content. Blog posts. Articles. Reviews. Comments. Social media. AI companies scrape the internet to train the next generation of models. Which means the next generation of AI is being trained on the output of the current generation. Each cycle loses information. Not randomly. It loses the rarest, most unusual, most creative parts first. The researchers call these the "tails of the distribution." The weird ideas. The unexpected perspectives. The things that made the internet feel human. Those disappear first. What remains is the average. The safe. The expected. The bland. Then the next generation trains on that. And loses more. And the next generation trains on that. And loses more. The researchers proved this is not a slow decline. Major degradation happens within just a few iterations. Even when some of the original human data is preserved. They tested it on large language models. On image generators. On statistical models. The pattern was the same every time. The output converges toward a narrow, flattened version of reality that looks nothing like the original data. The lead researcher put it plainly. "Large language models are like fire. A useful tool. But one that pollutes the environment." The pollution is invisible. You cannot see which sentence on the internet was written by a human and which was written by AI. Neither can the AI that is about to train on it. And once the tails are gone, they do not come back. The damage is irreversible. This is not a prediction anymore. It is a diagnosis. The internet you grew up on was built by humans writing things no algorithm would have written. Strange, personal, imperfect, alive. That internet is being diluted. One generation of AI at a time. And the models trained on what remains are learning a smaller and smaller version of the world. Model Collapse is not a technical problem. It is a cultural one. The thing that made the internet worth reading is the thing that disappears first.
1,112
6,230
1,370
17,422
2,313,956
David Fombella retweeted
"La IA ya está por delante de nosotros en la mayoría de tareas cognitivas." Geoffrey Hinton no lo dice como hipótesis. Lo dice como el hombre que construyó los sistemas que lo demuestran, y que luego dejó Google para advertirle al mundo. Su conferencia de 47 minutos no es un análisis del futuro. Es un diagnóstico del presente. Lo que más incomoda no es lo que dice sobre la IA. Es lo que dice sobre nosotros. La mayoría abre Claude, escribe algo y cierra la pestaña. Creen que están usando IA. Están usando el 10%. Hinton lleva años estudiando cómo aprenden las máquinas. Yo estudié su conferencia completa para entender cómo deberíamos usarlas nosotros. El resultado: un sistema de 18 pasos para usar Claude de verdad, con prompts listos para copiar y pegar hoy mismo. Lo tienes en el post de abajo.
7
47
175
37,264
I've just launched aifootball.co. As far as I know, it's the world's first dedicated marketplace for football AI tools. Skills, MCPs, and APIs. All in one place. 15 tools live today. Open to any developer who wants to list theirs.
14
46
1
467
48,294
David Fombella retweeted
Con lo guapo que había empezado…
25
2
169
4,274
Hamás (los amigos de Pedro Sánchez), roban los camiones con ayuda humanitaria. De esto que dice la cerda de Irene Montero ?
187
5,054
115
12,814
343,623
💣 Microsoft acaba de publicar un estudio donde determina los empleos más susceptibles de ser reemplazados por IA generativa (primera imagen) y los más seguros (segunda imagen) Os recomiendo que lo leáis, especialmente los jóvenes.
Readers added context they thought people might want to know
En este estudio no se mencionan los empleos "más susceptibles" de ser reemplazados por la IA, si no aquellos donde esta tiene mayor aplicabilidad, y viceversa en la imagen del siguiente tuit. Es una diferencia importante al contexto del tuit y cómo puede ser interpretado. arxiv.org/abs/2507.07935
338
2,587
757
11,296
2,240,787
⚽📊 Weekly loadings and recovery strategies for single and double game weeks journal.aspetar.com/en/archi…
1
47
340
35,052
David Fombella retweeted
¡Joder con Vito Quiles!
174
668
51
3,023
195,815
⚽📊Structuring the daily progression from return-to-run to full team integration sportperfsci.com/wp-content/…
35
186
10,310
¡DE AQUÍ NO SE MUEVE NI DIOS! #TalDíaComoHoy, 25 de abril de 1993, la sección del teniente Monterde se desplazó al norte del destacamento de #Jablanica en dirección a #Sarajevo. El convoy estaba formado por 5 blindados y 35 #legionarios.
25
268
12
933
65,705
David Fombella retweeted
Ball Striking - Lofted Pass. A Physical Literacy approach rather than the four-corner model. In terms of tactical, individual decisions would come under “think”. 🤔 welcome.
3
14
165
31,603
David Fombella retweeted
We’re hiring a Research Engineer at @Arsenal ⚽️ to help accelerate our applied research efforts! We’re building state-of-the-art models to understand football at scale, and a suite of ML-powered applications on top of them to drive tactical and strategic decisions.
212
749
266
6,781
820,181
⚽📊The Development of a Return to Performance Pathway Involving A Professional Soccer Player Returning From A Multi-Structural Knee Injury ijspt.scholasticahq.com/arti…
34
223
21,924
Será por tecnología para el rendimiento deportivo... (y, obviamente, no está todo).
11
104
2
571
66,962