David Sadigh

Entrepreneur at the intersection of luxury, data, and AI. Founder & CEO of DLG (Digital Luxury Group), co-host of The Luxury Society Podcast

[David Sadigh @DavidSadigh]

What’s in the knowledge base?
49 documents, organized into 8 synthesis files:

🧠 Psychological profile — childhood wounds, attachment patterns, limiting beliefs, personality type, recurring dreams, identity

💑 Relationship dynamics — communication patterns, conflict mechanics, love languages, progress over time

👨‍👧‍👦 Family — parenting challenges, generational patterns, what I’m trying not to repeat, what I’m building

📊 Mood & health cycles — 5+ years of mood tracking, addiction patterns, what triggers spirals up and down, sleep, sport as medicine

📈 Personal growth — 10 years of annual resolutions, self-assessments, the gap between knowing and doing

🧭 Values & identity — life principles, role models, what actually makes me happy (spoiler: not what I chase when I’m stressed), funeral exercise, gratitude journals

❓ Open questions — unconfirmed hypotheses, things to explore with a professional, signals to watch
Every file separates facts from patterns from hypotheses. Nothing is taken at face value — not even my own self-assessments.

[David Sadigh @DavidSadigh]

Psychology and AI :

I spent 10 years journaling, tracking moods, therapy sessions, writing resolutions I'd forget by February. Thousands of pages scattered across docs and notebooks.

Then it clicked: what if I turned all of it into a system that actually learns from my own history?

So I took everything — therapy notes, mood data, reflections, coaching exercises, annual reviews — structured it into a knowledge base, and gave it to Claude as permanent context.

Now I have an AI coach that truly knows me. Not generic "practice gratitude" advice. Real continuity. It connects what my therapist said to what I wrote last week. It sees patterns I missed across years of data.

It's not a replacement for therapy.
It's the memory therapy never had — and the most honest mirror I've ever looked into.

Latest output? A structured pre-diagnostic for ADHD, cross-referenced with 5 years of mood tracking, my childhood school reports, and my son's own diagnosis. Ready to hand to a specialist. That would have taken months to compile manually.

Will share more on the approach if you are interested.

10 years of self-work, finally talking to each other.

[David Sadigh @DavidSadigh]

There was a real critical vulnerability in Moltbook’s backend that posed a serious security risk — and it was addressed.
However, the specific claims about large-scale public exposure of sensitive personal data (emails, login tokens, API keys) affecting 1.5M+ users are currently not independently confirmed in authoritative sources.

[David Sadigh @DavidSadigh]

@Osint613 @grok when was this video recorded? Provide as much info as possible

[David Sadigh @DavidSadigh]

@CdricLabrousse J’ai trouvé @DelphineMinoui factuelle et objective sur France Inter hier…

[David Sadigh @DavidSadigh]

@sadokbenam @Arab_Intel Vous m’avez vu ici exprimer une opinion la-dessus ? Vous devriez relire je corrigeais juste une tournure factuelle ment erronée.

[David Sadigh @DavidSadigh]

Pour ceux que cela intéressent voici le résultat d’un agent de fact checking conçu par un expert dans l’analyse informationnelle, prenant en compte fact checking et biais cognitifs:

✅ Réponse : Grok est globalement plus précis que la formulation initiale, mais son verdict reste “partiellement vrai / non corroboré” — et il ne contredit pas mon évaluation précédente sur le fond.
🎯 Confiance : 93%

Je vais être très rigoureux et comparer claim par claim, sans posture défensive.

⸻

📌 Claim (tel qu’il circule)

« Channel 13 rapporte que, selon les renseignements israéliens, environ 500 membres des forces de sécurité iraniennes auraient été tués lors des manifestations. »

⸻

📚 Evidence (ce que Grok ajoute correctement)

Grok apporte un élément clé que tu n’avais pas explicitement fourni :

Channel 13 aurait parlé de “אנשי המשטר” (anshei ha-mishtar)
→ littéralement “membres du régime”, pas “forces de sécurité” au sens strict.

👉 C’est crucial.

Sur ce point, Grok a raison :
•✔️ Oui, Channel 13 a bien diffusé un sujet disant que
le cabinet israélien a été briefé par les renseignements israéliens
•✔️ Le chiffre ~500 est bien attribué à l’évaluation israélienne
•❌ Mais le terme utilisé est volontairement vague

⸻

🧠 Analysis (où se situe exactement la divergence)

1️⃣ Problème sémantique central (le nœud de l’erreur)

TermeSens
Forces de sécuritéPolice, IRGC, Basij, armée, forces anti-émeutes
“Membres du régime”Catégorie politique, pas militaire :peut inclure miliciens, cadres locaux, responsables du parti, agents, informateurs, parfois policiers

➡️ Traduire ou reformuler “אנשי המשטר” en “forces de sécurité” est une extrapolation.

👉 Donc :
•La version française initiale est trop spécifique
•Grok a raison de dire : “slight stretch”

⸻

2️⃣ Statut épistémique du chiffre “500”

Même Grok le reconnaît clairement :
•❌ Aucune source indépendante (ONU, HRANA, Amnesty, Reuters, AP)
ne confirme 500 morts côté forces de sécurité
•Les chiffres indépendants donnent plutôt :
•~40–50 forces de sécurité (HRANA)
•~100+ selon Téhéran (chiffres intéressés)
•Le chiffre 500 existe ailleurs, mais pour :
•morts totales
•ou autres contextes (frappes, 2025, etc.)

➡️ Donc ce “500” est :
•✔️ une estimation de renseignement israélien
•❌ pas un fait établi
•❌ pas vérifiable indépendamment

⸻

3️⃣ Nature du message de Channel 13

C’est un point que Grok explicite bien :
•Channel 13 ne fait pas une assertion factuelle neutre
•Elle rapporte une évaluation sécuritaire israélienne
•Dans un contexte de :
•dissuasion
•signal stratégique
•lecture de vulnérabilité du régime iranien

➡️ On est dans :

Intelligence assessment ≠ fait vérifié

⸻

🎯 Verdict comparé (clair et net)

🔹 Sur la phrase telle que formulée par l’utilisateur

« 500 membres des forces de sécurité iraniennes tués »

❌ Incorrect / trompeur

⸻

🔹 Sur la version corrigée par Grok

« Channel 13 rapporte qu’Israël estime ~500 membres du régime tués »

✔️ Exact comme citation médiatique
❌ Non vérifié factuellement
⚠️ Fortement dépendant d’une source intéressée

➡️ Verdict correct : “Partiellement vrai / contesté”

⸻

🚩 Flags – Biais & mécanismes d’influence

BIAIS_COGNITIFS_POTENTIELS
•Glissement sémantique : régime → sécurité
•Biais d’autorité : “renseignements israéliens” = crédibilité perçue
•Biais de cadrage : accent sur pertes du régime, pas des civils

MECANISMES_D_INFLUENCE_POTENTIELS
•Autorité : briefing du cabinet
•Rareté / urgence : contexte insurrectionnel
•Narrative stratégique : fragilité du régime iranien

⸻

Cc @Malbrunot

[David Sadigh @DavidSadigh]

J’ai vérifié et non: la formulation devrait plutôt être « La chaîne israélienne Channel 13 rapporte que, selon une évaluation des renseignements israéliens, environ 500 “membres du régime iranien” auraient été tués lors des troubles, une estimation non corroborée par des sources indépendantes et utilisant une catégorie large et ambiguë. » cc @Malbrunot

[David Sadigh @DavidSadigh]

Not sure weeks away, meeting with President Trump, Secretary of State Marco Rubio, Defense Secretary Pete Hegseth and Joint Chiefs Chair Gen. Dan Caine is planned for Tuesday. Memos are being sent to agencies
seeking their input on specific responses to the Iran situation—including potential military targets and economic options— ahead of the Trump meeting according to @WSJ

Understand they don’t have any aircraft carrier in the region but this won’t take them more than 15 days to bring back from Latin America.

End of January sounds doable, no?

[David Sadigh @DavidSadigh]

#iran “Memos are being sent to agencies seeking their input on specific responses to the Iran situation—including potential military targets and economic options—ahead of the Trump meeting, according to some of the officials.” - important meeting with @realDonaldTrump on Tuesday 🤞💪🏼🧿

[David Sadigh @DavidSadigh]

“One option under discussion is the possibility of the U.S. sending terminals of Starlink, a satellite-based internet service owned by Elon Musk, into Iran for the first time during the Trump administration, officials said, which could help protesters skirt a recent internet shutdown in the country.” The story of how @elonmusk helped in making this revolution a success will be thought across all Iranian schools 🤞💪🏼

[David Sadigh @DavidSadigh]

@thecurioustales @grok summarize this technique step by step in French

[David Sadigh @DavidSadigh]

@JosetteCaruso_ @Polymarket @rawsalerts @grok explain this

[David Sadigh @DavidSadigh]

@zephyr_z9 Former Sensetime senior engineers - they were already quite advanced across a bunch of uses cases including road traffic analysis and forecasting many years ago

[David Sadigh @DavidSadigh]

Capital Account: As global fertility rates drop, two economists make the case for humans https://t.co/3RgcaYJr5F

[David Sadigh @DavidSadigh]

What does Meta’s new "free speech" strategy mean for luxury advertisers?

Recently, Meta has taken a bold new direction, making calculated moves to align itself with the latest U.S. presidential administration. From restructuring leadership to prioritizing "free speech" over moderation, the company has captured significant attention—but what does this mean for luxury brands?

Luxury advertising relies on controlled, premium environments to protect brand equity and align with consumer expectations. Meta's latest strategy raises critical questions:

Will Meta put brand safety at risk, and how might luxury advertisers respond?
Will these moves alienate luxury brands that demand qualitative environments for storytelling?
How can Meta reassure its premium advertisers while pursuing its broader political ambitions?
Who can potentially benefit from this move? TikTok? Youtube? Teads?

What’s your take on this?

We’ll analyze these developments and their impact on the luxury landscape—spanning advertisers, platforms, and the consumer experience. More to come on Luxury Society.

[David Sadigh @DavidSadigh]

Mexico is sitting on more than half a billion litres of tequila in inventory, almost as much as its annual production, as the fast-growing industry reckons with slowing demand and the prospect of tariffs on exports to the US under DT. #tequila #slow https://t.co/dVQY5p46Ci

[David Sadigh @DavidSadigh]

Gen AI Fail: Some creative folks are boasting about their achievements with Gen AI in advertising. Let’s just hope their clients are blind, because the output is far from impressive…
#genaifail

[David Sadigh @DavidSadigh]

Bookmark this 👌- The 2025 AI Engineering Reading List - Latent Space #AI # https://t.co/vfJveTJrAC