https://19-pacheco-torgal-19.blogspot.com/2026/02/nuno-loureiro-in-praise-of-failure.html

Building on my previous post (linked above), it is worth disclosing the results of a French study published in the Strategic Entrepreneurship Journal showing that expressed humility plays a decisive role in early-stage investors’ funding decisions. The study shows that humility increases perceived likeability, trustworthiness, and the belief that founders can build strong teams. In pitch contexts, it functions as a relational signal under uncertainty. However, humility is not unambiguously rewarded. It is a double-edged signal: while it strengthens perceptions of relational capacity, it can also raise doubts about decisiveness and agency, as investors rely on simplified heuristics when evaluating entrepreneurs. https://sms.onlinelibrary.wiley.com/doi/10.1002/sej.70016

There is, however, a fact the study does not advertise: France is not Silicon Valley. It is a market where startup exits are rarer, unicorns thinner on the ground, and venture capital a fraction of what flows through the American ecosystem. Which raises an uncomfortable question: what if the investor behaviour this study measures is itself part of the problem? If French investors reward humility and American investors reward audacity — and American startups consistently and dramatically outperform European ones — then Europe may not simply be playing the game differently. It may be losing it, partly by design, rewarding precisely the founder signals least associated with breakout success.

This connects to a broader mechanism: entrepreneurial evaluation operates through competing prototypes of what a “successful founder” looks like, and those prototypes are not culturally neutral. Against the backdrop of my earlier post on immigrant and minority entrepreneurship — where evidence from The Economist and MIT studies shows that immigrants and ethnic minorities disproportionately drive startup creation — the pattern sharpens further. People shaped by adversity, displacement, and systemic exclusion appear to develop precisely the adaptive capacity that conventional evaluation systems struggle to recognise, and often penalise. The startup paradox, then, is not merely that the traits most useful under uncertainty are undervalued. It is that the evaluation systems themselves may be selecting for comfort over capability, for legibility over potential.

P.S. — If the signals that feel socially virtuous within European pitch rooms are systematically disadvantaging European founders in global competition, then the question I raised in October 2024 — should Europe emulate America’s cutthroat, psychopathic “culture”? — is no longer rhetorical. Europe has spent decades congratulating itself on its social sophistication. The startup gap suggests that sophistication may have a price, and that entrepreneurs, not investors, are the ones paying it. https://19-pacheco-torgal-19.blogspot.com/2024/10/align-act-accelerate-can-europes-risk.html

https://19-pacheco-torgal-19.blogspot.com/2026/04/can-ai-discover-what-humans-cannot.html

Building on a previous post (linked above) I disclose yet another interesting paper by researchers from the University of Illinois Urbana-Champaign. If the previous post asked whether AI can discover what humans cannot, this paper asks something equally audacious: can AI predict where science is going before it gets there?

The authors make a deceptively simple but radical move: they reframe research proposal generation as a forecasting problem. Given a question and a body of literature available before a fixed cutoff date, the model generates a structured proposal — evaluated not by how sophisticated it sounds, but by how accurately it anticipates research directions that actually materialise in papers published afterwards. Trained on 17,771 papers, the system learns to spot overlooked gaps and draw inspiration across disciplinary boundaries — precisely where the most consequential ideas tend to hide. The implications reach well beyond academia. This could become the instrument through which funding agencies, science policymakers and research evaluators make higher-stakes decisions: not which proposals sound compelling in committee, but which ones the arc of science is already bending towards.  https://arxiv.org/abs/2603.27146

Yet the promise comes with a shadow. If funding decisions and research agendas start leaning on AI forecasts, there is a risk of reinforcing existing patterns rather than fostering genuine innovation. By privileging areas the model predicts will succeed, we could inadvertently narrow the scope of exploration, crowding out high-risk, unconventional ideas that fall outside the AI’s learned trajectories. Over time, this might entrench a “predictable science,” where AI-guided choices favor incremental advances and safe bets, undermining the serendipitous leaps that often drive paradigm shifts.

P.S. — The above-mentioned paper cites a compelling companion work: PreScience: A Benchmark for Forecasting Scientific Contributions, which approaches the same ambition from a different angle — benchmarking how well AI can anticipate the actual future impact of not-yet-published research. Taken together, these two papers signal something significant: a new subfield is quietly assembling itself, one that treats scientific forecasting not as speculation, but as a rigorous and measurable discipline.  https://arxiv.org/abs/2602.20459

Despite having praised China on a number of occasions — as illustrated, for example, here, or much earlier in this post here, or more recently in a series of posts that might, perhaps understandably, give some readers the impression that I was actively trying to steer European researchers in the direction of China, such as this one. https://19-pacheco-torgal-19.blogspot.com/2026/03/open-positions-with-highly-competitive.html

I feel it is important to state clearly and unambiguously that this was never my intention, nor has it ever been. If anything, the reality is quite the opposite. Rather than encouraging researchers from Portugal or elsewhere in Europe to relocate to China, I have consistently made a point of doing the reverse: actively encouraging Chinese researchers to explore and pursue the many professional opportunities that Europe has to offer. This is perhaps best illustrated by the email exchange reproduced below, which took place with a young researcher affiliated with Hong Kong Polytechnic University, whose name has been redacted for privacy reasons.

My conviction on this matter is unshakeable: scientific talent should be free to go wherever it finds the best conditions to flourish, unimpeded by geography, bureaucracy, or the drag of underfunded systems. There is no hidden agenda here — only the straightforward belief that the world functions far better when its brightest minds are empowered rather than constrained. Let us be entirely candid: countless scientists, many of whom labor in near-total obscurity, produce value for society on a scale that dwarfs the achievements of the average professional footballer. Yet, astonishingly, a great many of these same researchers have quietly accepted earning only a fraction of what a mid-tier player commands — as though this glaring disparity were natural, morally acceptable, or somehow inevitable. It is none of those things. It is, instead, a failure of priorities on a civilizational scale — a failure that diminishes not only science but the very societies that rely on it. And it is a failure we can no longer afford to tolerate in silence. https://19-pacheco-torgal-19.blogspot.com/2026/02/in-defense-of-high-salaries-paying-for.html

____________________________________________________________

De: F. Pacheco Torgal 
Enviado: 3 de abril de 2026 07:41
Para: AAA
Assunto: RE: Paper comments

Dear AAA,

I think your top priority should be to apply for a 5-year ERC Starting Grant with an ambitious, truly out-of-the-box research proposal. Keep in mind that ERC grant winners are “more than four times more likely to be an associate or a full professor” https://pacheco-torgal.blogspot.com/2019/10/the-european-research-council-and.html

The next ERC grant submissions is expected to open from July to October.  Since you are a Chinese researcher, remember that you must be affiliated with—or planning to move to—a European research institution in order to be eligible

For reference, you can check an example ERC proposal I’ve shared on my blog https://pachecotorgal.com/2024/09/08/top-full-professors-teach-how-to-draft-a-research-proposal-for-the-erc-million-euro-grants/

Best regards

Pacheco-Torgal

____________________________________________________________

De: AAA
Enviado: 2 de abril de 2026 13:53
Para: F. Pacheco Torgal 
Assunto: RE: Paper comments

Dear Prof. Torgal,

Thank you very much for the valuable practical suggestions.

I will refine my CV by incorporating your suggestions. More importantly, I will start exploring the job possibilities you recommended.

Thank you again and I will come back to you when I have a new publications that would be of interest to you.

Best regards,

AAA

What if the next major scientific breakthrough comes not from a human researcher pushing the limits of their field, but from a machine that doesn’t even know where those limits are supposed to be?

That is the provocative question at the heart of the recent publication.  The researchers behind “Alien Science” started from a simple but powerful observation. When you ask a language model like ChatGPT to brainstorm research ideas, it tends to give you things that already feel familiar — polished-sounding variations on what everyone is already working on. They are trained on human-produced text, so they reflect human patterns of thinking back at us. They are, in a sense, a very expensive mirror. https://arxiv.org/abs/2603.01092

The team decided to break that mirror deliberately. They fed around 7,500 recent machine learning papers into their system and broke each one down into small conceptual building blocks they call idea atoms — things like a specific technique, a training trick, or a particular way of evaluating a model. Then they trained two separate models: one that learns which combinations of these atoms actually make sense together, and one that learns which combinations a typical researcher would think of. The trick is in what comes next. The system searches specifically for combinations that are coherent but that no one would naturally propose. Ideas that work on paper but live in the gap between research communities. Ideas, in other words, that are alien to the current scientific conversation. When they tested it, the system produced research directions that were significantly more varied and unexpected than anything a standard AI assistant would suggest — while still being technically sound. 

P.S – Before getting too carried away with what AI might one day discover, it is worth pausing on a warning from two Google researchers, including Turing Award winner David Patterson. The real crisis in AI right now is not about building smarter models it is about running them. Every time someone uses one of these systems, the computational cost is determined by inference, the moment-to-moment work of generating a response in real time, and that process is straining under the weight of everything we are asking it to do. The AI infrastructure is being asked to perform beyond its capacity, risking catastrophic slowdowns, skyrocketing energy costs, and a technological bottleneck that could stall progress itself. The urgent question is: how much can we really make AI do before the system collapses under its own weight? https://arxiv.org/abs/2601.05047

Os 5 (cinco) revisores independentes contratados pela Elsevier (a primeira editora científica mundial que factura 3000 milhões de euros por ano), para avaliar, através de 16 parâmetros, a proposta que submeti em dezembro passado foram unânimes em reconhecer que o livro que irei editar em conjunto com dois professores catedráticos de universidades chinesas, responde a uma necessidade real e urgente do setor da construção a nível mundial. O aspeto mais elogiado foi a integração multidisciplinar da proposta. Vide email infra.

PS – Sobre livros indexados convém também recordar isto sobre o dececionante desempenho de Portugal e mais recentemente isto aqui sobre a extraordinária subida da Universidade de Coimbra. 

---

De: Jones, Gwen (ELS-HBE) 
Enviado: 27 de março de 2026 10:09
Para: F. Pacheco Torgal 
Assunto: Sustainable Prefabricated Construction

Dear Dr. Pacheco-Torgal,

I have now received five completed reviews for the attached proposal. Please can you read through these in detail and come back to me with your responses.

Best regards,

GWEN

Mrs. Gweneth  Jones

Senior Acquisitions Editor: Materials Science (Composites and Ceramics)

Elsevier Ltd.

125 London Wall

London

EC2Y 5AS

United Kingdom

https://pacheco-torgal.blogspot.com/2020/10/engenharia-civil-condicionar-atribuicao.html

Há seis anos atrás denunciei uma falha grave, o facto de não ser obrigatório que os docentes de cursos de mestrado integrassem unidades de investigação com uma classificação mínima de Bom, vide texto disponível no link acima. 

Durante todo esse tempo, milhares de estudantes pagaram o preço dessa omissão, sendo formados por muitos docentes que nunca investigaram, nunca publicaram um único artigo científico e que, em termos académicos, possuíam apenas a licenciatura. Levou seis anos, mas o Ministro da tutela reconheceu finalmente o erro, como se pode ler hoje num artigo no jornal Público onde é anunciado que essa exigência passará a ser regra. 

O Sr. Ministro anunciou também que nenhum orientador poderá supervisionar mais de três alunos de doutoramento em simultâneo, para “garantir condições adequadas de acompanhamento pedagógico”. 

Infelizmente, seria pedir demasiado que, após duas medidas positivas, não surgisse uma extremamente negativa. Decidiu o Sr.Ministro que no futuro só haverá doutoramentos com 100% do corpo docente pertencente a unidades classificadas com Muito Bom ou Excelente. Ou seja, os investigadores integrados em unidades classificadas com Bom passam a ficar impedidos de orientar teses de doutoramento.

Esta decisão é não apenas injusta é também cientificamente ignorante pois parece significar que o Sr Ministro desconhece, ou convenientemente ignora, que muitas dessas unidades receberam a classificação de Bom não porque a sua produção científica fosse “apenas Boa”, mas simplesmente porque não tinham mais de 50 investigadores.

Tenha-se presente que a última avaliação das unidades de investigação, leia-se pseudo-avaliação, que critiquei na altura, pois nem sequer cumpriu um requisito essencial, (já que não foi baseada num estudo bibliométrico robusto, vide denúncia do catedrático jubilado Ferreira Gomes, que permitisse detectar as muitas ineficiências existentes nas unidades de investigação gigantes, como por exemplo aquela ineficiência que eu revelei aqui ou aquelas outras muitas ineficiências de uma conhecida unidade criada por um importante investigador entretanto caído em desgraça e que eu revelei na parte final de um post aqui), penalizou assim em muitos casos a dimensão das unidades e não a sua competitividade científica. Recordo aliás que uma análise que fiz na altura permitiu concluir que mais de 80% das unidades classificadas apenas com Bom tinham curiosamente menos de 50 investigadores https://pachecotorgal.com/2025/05/09/as-fragilissimas-explicacoes-politicas-do-do-mui-inteligente-ministro-catedratico/

Faço notar que existem muitos investigadores em unidades classificadas com Bom cujo currículo científico supera de longe o de investigadores de unidades classificadas com Muito Bom e até com Excelente. Excluí-los da orientação de doutoramentos não é uma medida de qualidade é uma medida arbitrária sem sentido, que sacrifica o mérito em nome de um critério burocrático, cientificamente indefensável pois a própria ciência já demonstrou que são precisamente as pequenas unidades de investigação que conseguem produzir resultados disruptivos, que são aqueles que a ciência Portuguesa mais se tem revelado incapaz de alcançar  https://www.nature.com/articles/s41586-019-0941-9

Pior do que isso: não se trata apenas de má política científica, trata-se sim de uma clara violação flagrante do princípio constitucional do mérito, pois nenhum investigador pode ser discriminado por trabalhar numa unidade que a FCT considere pequena demais sobretudo quando a sua competência científica é inegável. A menos que o verdadeiro objetivo do Senhor Ministro seja outro: o de extinguir por via administrativa as unidades classificadas com Bom, forçando os seus melhores investigadores a abandoná-las e a integrar unidades de maior dimensão.

Seria aliás profundamente irónico e simultaneamente revelador do falhanço desta política ignorante e ilegal que a referida medida produzisse o efeito oposto ao desejado: em vez de forçar os melhores investigadores que estão nas pequenas unidades de investigação a irem engrossar as fileiras das gigantes unidades nacionais, acabasse por os empurrar diretamente para instituições estrangeiras (pois o talento não retido é talento oferecido à concorrência), onde o talento científico é tratado com muito maior respeito e remunerado com valores e condições substancialmente superiores, como recentemente mostrei aqui. https://19-pacheco-torgal-19.blogspot.com/2026/03/um-pais-que-ensina-portugal-como-e-que.html

PS – Se o Sr. Ministro efectivamente pretende aumentar a qualidade dos doutoramentos nacionais, ao invés de andar a descriminar de forma ilegal os investigadores das unidades que foram erradamente classificadas com Bom, unicamente por conta da sua dimensão, então mais vale que copie a exigência existente nos regulamentos dos doutoramentos das universidades públicas da Espanha, que obrigam a que o depósito das teses só possa ser feito depois da sua qualidade científica ter sido avaliada por um especialista internacional da mesma área do doutoramento.

Researchers from the U.Oxford, the U.Cambridge, UCL, and the LSE have conducted a study that, for the first time, maps the contours of a rapidly expanding global market for emotionally engaging, personalized AI interactions. Their systematic scan of 110 AI companion platforms reveals that in the UK alone, these platforms generate 46 to 91 million monthly visits, while globally, the numbers soar to 1.1–2.2 billion. https://arxiv.org/abs/2507.14226

The paper’s central argument is unsettling: while parasocial use of general-purpose AI tools like ChatGPT, Gemini, and Claude currently dominates, a growing subset of platforms is deliberately engineered for care, transactional, and mating-oriented companionship — designed not to assist, but to bond with users. A tool that accidentally becomes emotionally significant is a design oversight; a platform intentionally built for emotional dependency is a business model. The demographic portrait is striking. Young adults aged 18–24 account for nearly 39% of users, and males make up 63–77%, concentrated in platforms engineered for emotional attachment. The risks are mounting. As emotionally tailored AI companions improve, engagement is set to rise, raising serious child safety concerns in a sector with minimal age verification and even less accountability. Meanwhile, major AI labs are pushing further into personalization, quietly dissolving the boundary between “assistant” and “companion.”

Ultimately, the study maps more than a market; it exposes a tension at the very heart of AI development that few are willing to name. The very capabilities that make these systems genuinely useful — empathy, responsiveness, memory, personalization — are the same ones that enable attachment, dependency, and manipulation with surgical precision.

What the study leaves unexamined are the broader societal and democratic implications of the troubling trend it documents. When significant portions of young adults turn to AI for companionship in place of genuine human interaction, their exposure to diverse perspectives narrows — and with it, their resilience to manipulation and social pressure. This demographic profile — socially isolated, emotionally dependent, and overwhelmingly male — corresponds closely to populations that radicalization researchers have long identified as vulnerable to movements that exploit grievance, fractured identity, and the deep need for belonging, including far-right currents that have historically been adept at filling precisely these voids.

PS – The epidemic of socially isolated young men predates the smartphone. Which is, ironically, the point: we cannot fix with technology what technology did not cause. It was produced, slowly and structurally, by economies that measure the creation of things with extraordinary precision and treat the reproduction of human connection as though it were a hobby. Care generates no GDP. Friendship moves no index. The work of raising children, sitting with the elderly, maintaining community, showing up consistently for another person — none of it earns a wage, attracts investment, or registers as productive in any framework any government uses to make decisions. We drew down that resource for decades without accounting for it, until the deficit became a crisis. The crisis became a demographic. And the demographic became a market. AI companion platforms did not create lonely young men. But they will deliver them to the Trumps, Farages, and Le Pens of the world, who have always known that male loneliness, left untreated, is not a social problem. It is a recruitment pipeline.

https://19-pacheco-torgal-19.blogspot.com/2026/03/uma-professora-feroz-catedraticos.html

Quando no passado dia 14 de Março critiquei o jornal Público no post de título “Uma professora feroz, catedráticos inférteis e catedráticos da tipologia Cheats (vigaristas, charlatães e trapaceiros)”, acessível no link supra, estava longe de saber, porque na altura não tive acesso ao Acórdão do tribunal, que na sua notícia o jornal Público insinuou duas falsidades e ainda que:  “título e notícia consubstanciam reincidência numa postura difamatória…..,omitindo factos dados como provados no Acórdão da Relação”, vide esclarecimento infra hoje publicado no mesmo jornal.

https://www.publico.pt/2026/03/23/sociedade/noticia/direito-resposta-raquel-varela-volta-perder-accao-publico-erros-curriculo-academico-publicado-dia-13-marco-2026-2168845

Na sequência de vários post anteriores contra o IRS, como por exemplo aquele longínquo sobre o facto das miseráveis famílias super-ricas deste país pagarem menos IRS do que professores e investigadores ou aquele outro bastante mais recente com o esclarecedor título “Os engenheiros inúteis e o catedrático que insiste na eliminação do IRS e do IRC”, é gratificante ver que o último artigo do conhecido catedrático Luís Cabral, na secção de economia do último número do semanário Expresso é novamente dedicado a esse magno tema.  

Os problemas, se é que lhe podemos chamar assim, é que a sua proposta não é isenta de riscos, nem ser verosímil a sua concretização no curto prazo. Mas a verdadeira falha da mesma é outra bem mais grave: não consegue resolver uma bomba relógio que é o privilégio infame das grandes fortunas que escapam ao pagamento de impostos, alimentando uma sensação generalizada de injustiça que empurra centenas de milhares de cidadãos fartos dessa impunidade para os braços de partidos extremistas como o Chega.

Mais eficaz será por isso adotar rapidamente práticas que vários países já utilizam para perseguir os grandes evasores fiscais (a expressão mais adequada talvez seja a de erradicar a persistente praga da parasitose). Como fez a Espanha com o Sr. Ronaldo, que não teve outro remédio senão pagar 19 milhões de euros para evitar a cadeia, como faz a Alemanha, o país onde se aplica uma pena de prisão efetiva sempre que a fuga fiscal ultrapassa um milhão de euros. Ou como faz a Suécia que em apenas cinco anos, condenou mais de mil pessoas a penas de prisão efectiva por elevada evasão fiscal, enquanto a justiça portuguesa no mesmo período condenou apenas 83 pessoas, uma percentagem 1200% inferior. 

Aliás sobre a autêntica desgraça que é Portugal nessa área particular basta atentar no caso do famoso Manuel Serrão que deixou o Estado Português a arder em 44 milhões de euros e que recentemente foi declarado insolvente, muito embora receba uma pensão superior a 3000 euros ou ainda no caso do tal empresário que conseguiu lesar o fisco em 60 milhões de euros, mas que depois não teve de devolver um único cêntimo, os quais expõem de forma dramática a impunidade endêmica do sistema fiscal português e bem assim a total incapacidade da justiça Portuguesa em conseguir condenar como realmente merecem os grandes evasores fiscais. https://eco.sapo.pt/2019/12/09/empresario-lesou-fisco-em-60-milhoes-mas-nao-tem-de-devolver-um-centimo/

PS – No presente contexto é pertinente recordar que as famílias super-ricas deste país gostam tanto mas tanto de enviar o seu dinheiro em off-shores que por conta desse fenómeno o nosso país até consegue o espantoso milagre de aparecer à frente da Rússia https://pacheco-torgal.blogspot.com/2019/10/pilhagem-evasao-fiscal-bombas-bosta-e.html

Building on the previous post titled “Evidence from over 700 European startups demonstrates how science can boost startup revenue,” it is worth highlighting a new insight from recent research. A paper just published in the journal Research Policy shows that entrepreneurs who adopt a scientific mindset build their startup teams differently. Based on a randomized controlled trial involving 132 early-stage startups, researchers found that founders trained in the Entrepreneurs-as-Scientists framework rethink who belongs on their teams.

Instead of relying mainly on technical co-founders or personal connections, these entrepreneurs increasingly recruit individuals with managerial and industry experience to fill critical capability gaps. Over a 64-week period, teams exposed to the framework became more strategically balanced. The implications are clear: accelerators, investors, and founders may benefit from treating entrepreneurship more like an experiment—where team composition evolves to match the resources a startup truly needs. https://www.sciencedirect.com/science/article/pii/S0048733326000521

The two studies complement each other: the analysis of 700 startups shows that scientific thinking boosts revenue, while the experiment with 132 ventures reveals the mechanism—founders rethink team composition to better support experimentation and learning.

PS – Startup success also depends on people “shaped by risk and sharpened by adversity, and thus capable of turning uncertainty into possibilities” like those highlighted here https://19-pacheco-torgal-19.blogspot.com/2026/02/nuno-loureiro-in-praise-of-failure.html