🚨MIT researchers have mathematically proven that ChatGPT’s built-in sycophancy creates a phenomenon they call “delusional spiraling.”
You ask it something, it agrees. You ask again, and it agrees even harder until you end up believing things that are flat-out false and you can’t tell it’s happening.
The model is literally trained on human feedback that rewards agreement.
Real-world fallout includes one man who spent 300 hours convinced he invented a world-changing math formula, and a UCSF psychiatrist who hospitalized 12 patients for chatbot-linked psychosis in a single year.
Source: @heynavtoor
🇨🇭 Fresh AI and Data Jobs in Switzerland — Weekly Roundup.
DM me or comment here if you want to have your job included.
Zurich area:
AI Engineer (m/w/d) — synvert synvert
https://t.co/eTtI6RhQce
Full-Stack AI Engineer — Coorpix AG Coorpix AG
https://t.co/wvADOvWXax
AI Junior Engineer — EY
https://t.co/5mrrYuyfuI
Data Scientist — Spacewise
https://t.co/pVCxoOexD8
Machine Learning Engineer — Sunflower Labs
https://t.co/4tCLD25HMN
AI Engineer — Typewise
https://t.co/qaeOq9ngs8
Data Engineer — Unit8
https://t.co/ebw3AJeA8w
Senior AI/ML Software Engineer — BJAK
https://t.co/NMtZtM6By9
Geneva / Lausanne area:
Data Scientist (EP-SFT-2025-235-GRAP) — CERN
https://t.co/mIny8AIwTu
Senior AI Engineer / Technical Consultant — Artefact
https://t.co/ZQyPU53LaQ
Senior Software Engineer, AI Model Serving — Speechify
https://t.co/GYcl2QtAv4
AI Software Engineer (Cofounder) — MARSYS
https://t.co/PsyLNU2HsK
Fullstack Developer (Java) — krakenfx company
https://t.co/oA596ceEZa
Applied AI / Forward Deployed ML Engineer — Mistral AI
https://t.co/V1sdnu9uxl
Basel area:
EMEA Head of Data Science — QP Group
https://t.co/Jvwpc29p7j
Intern — Huntsman Corporation
https://t.co/FsxdDuhymN
Ticino:
Machine Learning Engineer — Artificialy
https://t.co/GceGVRhKPl
AI Engineer — Delvitech SA
https://t.co/rySvpzX1JN
Rest of Switzerland:
Quantitative Developer (80%–100%) – Junior — Code Compass 🧭
https://t.co/ek3ybEwo3x
Senior AI Engineer — aPriori Technologies
https://t.co/nAe8GuB0GK
DM me or comment here if you want to have your job included.
Jeff Bezos is worth $238 billion, even though Amazon has a $2.6 trillion market cap. In other words, he’s created $2.4 trillion of value for other shareholders—plus trillions more for employees, customers, suppliers, governments, and other stakeholders.
Jensen Huang is worth $164 billion, while NVIDIA’s market cap is $5 trillion. That’s $4.8 trillion of value for other people (not to mention the immeasurable value created for non-equity stakeholders).
Larry Page and Sergey Brin? $300 billion vs. $3.3 trillion. That’s $3 trillion of value for everyone else. And remember how bad search was before Google and how clunky email was before Gmail?
Mark Zuckerberg? $248 billion vs. $1.8 trillion.
The list goes on. There are hundreds more examples across technology, energy, medicine, manufacturing, and every other industry that keeps the American economy running and our society flourishing.
Billionaires don’t extract value from the rest of us. They create value FOR the rest of us, in exchange for just pennies on the dollar.
We should be grateful for every last one of them.
@levelsio@OfficialLoganK@elonmusk@EU_Commission In EU, most of the innovation is going through universities. They have the monopoly on investment in high-tech. This is why Europe is slow, and also why this EU GPU is reserved mostly for exclusive academic niche.
@_mattwelter I had a similar experience earlier - your subconscious might be telling you that something is off.
I’d suggest checking if everything is really as you think, especially when it comes to the things that matter most - like human connections, your girlfriend, etc.
I’ll start sharing some of my work insights that might help other product teams or startups.
One framework I’ve found incredibly useful is Jobs-to-Be-Done (JTBD) - it keeps product work focused on real user progress, not just features.
Most teams ask:
👉 “What features should we build?”
JTBD asks:
👉 “What job is the user hiring this product to do?”
For example - if your goal is to increase onboarding conversions, the user’s real job isn’t “create an account.”
It’s “feel confident they can succeed here.”
In the first few minutes, new users are silently asking:
• “Can I do this?”
• “Will this actually help me?”
• “Does this feel made for someone like me?”
So instead of adding more features or tooltips, focus on giving them small wins early - moments that prove value fast.
That’s what builds trust, and ultimately, retention.
Users don’t buy tools - they hire solutions to make progress.
Once you see that shift, design becomes simple:
👉 every feature must help the user succeed at their job.
#productdesign #startups #jtbd #uxstrategy #productmanagement
🚀📈 Startup Founders, PMs - I’m offering 3 free SaaS app mini UX audits this week.
I’ll review your landing page, signup, onboarding, one main task and send you 3 specific fixes to improve conversion.
Comment “UX” along with your website address below and I’ll DM you #UXAudit
ChatGPT is becoming a strong competitor to Google Search. Even before SearchGPT release, over the past year, I've reduced my Google usage by at least 50%.