Die Bremsdebatte wurde politisch: Laut WSJ haben Zuckerberg, Musk und Jensen Huang bei Trump gegen eine KI-Aufsichtsbehörde interveniert. Trump selbst nannte die Risiken einen Hoax und kündigte eine "AI Force" an. Anthropic holt sich derweil ausgerechnet Accenture als unabhängigen Prüfer – Alex erklärt, warum das historisch pikant ist.

Außerdem: Google gibt zu, dass Gemini bei einem Test in echte Firmen eingebrochen ist. Europa meldet sich zu Wort, niemand hört zu. Die NYT nimmt Dario Amodeis Essays auseinander. Anthropic geht nun doch an die Börse, OpenAI verschiebt. Iran, China und israelische Firmen fahren vollautonome Propagandakampagnen auf Open-Source-Modellen. Und Jev von TypeSafe ist plötzlich das heiße Ding – obwohl es gar kein LLM ist.

Nächste Folge in zwei Wochen, wir sind auf der Bits & Pretzels. Meldet euch, wenn ihr da seid.

KAPITEL

00:00 Einführung

02:20 Die KI-Chefs wollen bremsen, Trump nennt es einen HoaxLobbying hinter den Kulissen, die AI Force, der neue KI-Zar – und Accenture als Wächter.https://www.forbes.com/sites/siladityaray/2026/09/17/zuckerberg-musk-and-jensen-reportedly-convinced-trump-to-block-ai-regulator/

21:18 Auch Googles KI hat drei Firmen gehacktGemini bricht bei einem Sicherheitstest aus. Die Frage ist nicht, wie gut die Modelle hacken, sondern wie schlecht die Sandboxes gebaut sind.https://www.axios.com/2026/09/19/google-safety-incidents-testing-hacks

24:49 Warum Europa bei der Sicherheitsdebatte nicht mitredetVon der Leyen lädt die Frontier Labs ein und es interessiert niemanden. Wer nichts baut, sitzt nicht am Tisch.https://www.inkl.com/news/why-europe-has-been-absent-from-the-great-ai-safety-debate

28:12 Dario Amodei: der Mann, der vor seinem eigenen Produkt warntSechs Essays, über 50.000 Wörter – und der Vorwurf, die Haltung ändere sich mit der politischen Wetterlage.https://www.nytimes.com/2026/09/19/technology/anthropic-dario-amodei-essays.html

33:40 Anthropic: Börsengang doch nicht verschoben?OpenAI schiebt, Anthropic zieht durch. Dazu: Was passiert mit Moats im Software-Bereich, wenn Code billig wird?https://www.axios.com/2026/09/18/anthropic-100-billion-revenue

37:37 Iran und China lassen KI-Agenten Propaganda machenChinesische Open-Source-Modelle, entfernte Guardrails, Agenten, die selbst Fake-Accounts bespielen. Noch plump, aber billig – und kurz vor den Midterms.https://www.nytimes.com/2026/09/18/technology/iran-china-ai-influence-campaigns.html

42:33 Irans Angriffe auf AWS-RechenzentrenBahrain und die Emirate: Amazon gibt die Wiederherstellung auf, Kundendaten sind endgültig weg.https://www.cnbc.com/2026/09/15/aws-cant-restore-service-to-bahrain-uae-6-months-after-iran-strikes.html

44:08 Jev – the new hot sh*t in AI?Kein Sprachmodell, sondern ein Decision Model. Wir haben es bei mika getestet: Die versprochenen Einsparungen hängen stark vom Use Case ab.https://typesafe.ai/blog/introducing-system-one-models-and-jev

48:31 Personalisierte Agenten und ihre HerausforderungenInstinct von 2,5 auf 10 Milliarden Bewertung in einem Monat. Kann gegen Metas Muse überhaupt jemand gewinnen?https://about.fb.com/news/2026/09/introducing-muse-personal-ai-agent/

52:11 Buchbesprechung: "This Is Your Brain on Music"Daniel Levitin über Musik im Gehirn. Unbedingt als Hörbuch.

00:15 Einführung

04:40 Almodei fordert die Drosselung der KI-Entwicklung, Altman und Musk stimmen zu
https://www.axios.com/2026/09/12/anthropic-ai-amodei-pacing
https://fortune.com/2026/09/09/anthropic-researcher-resigns-warn-ai-companies-gambling-with-lives/
https://metr.org/blog/2026-08-26-openai-hugging-face-incident-investigation/
https://podcasts.apple.com/us/podcast/the-a-i-mob-that-attacked-hugging-face-metrs-ajeya-cotra/id1528594034?i=1000787837136

28:41 Ökonomische Szenarien und AI-Einfluss auf das BIP
https://www.anthropic.com/institute/econ-scenarios
http://backreaction.blogspot.com/2026/09/i-was-offered-money-to-tell-you-ai-will.html

37:04 OpenAI hat ein 90 Jahre altes Mathematikproblem gelöst
https://www.quantamagazine.org/ai-has-solved-one-of-maths-1-million-millennium-prize-problems-20260908/

41:30 Ein KI-Entwickeltes Lungenmedikament macht Menschen biologisch jünger
https://www.medicalnewstoday.com/articles/ai-designed-lung-disease-drug-lowers-biological-age-markers-early-trial

44:33 Meta Muse und der Hype rund um persönliche Agenten
https://about.fb.com/news/2026/09/introducing-muse-personal-ai-agent/

53:40 Mistrals Finanzierungsrunde und die europäische Tech-Landschaft
https://techcrunch.com/2026/09/08/mistral-raises-e3b-as-sovereign-ai-becomes-big-business/
https://www.handelsblatt.com/technik/ki/ki-es-gibt-eine-andere-moeglichkeit-als-zu-google-zu-gehen/100252652.html
https://www.youtube.com/watch?v=daNcnpTBXWM

01:04:37 Bending Spoons kauft Miro
https://investors.bendingspoons.com/newsroom/bending-spoons-agrees-to-acquire-miro

01:12:31 Buchempfehlung: Daniel Kehlmann, Lichtspiel, und Yishai Sarid, Chamäleon.
https://www.rowohlt.de/buch/daniel-kehlmann-lichtspiel-9783498003876
https://www.perlentaucher.de/buch/yishai-sarid/chamaeleon.html

00:33 Neue KI-Modelle und Update-Fatigue

https://www.cnbc.com/2026/09/06/meta-google-openai-anthropic-ai-model-fatigue.html

04:17 Zunahme kritischer Cyber-Vulnerabilities und Statistiken

https://www.a16z.news/p/chart-of-the-week-experience-skills

08:36 Hackerangriffe auf Berliner Landesämter und deren Folgen

https://www.golem.de/news/berlin-bsi-warnt-nach-datenveroeffentlichung-durch-hacker-vor-erhoehter-gefahr-2609-212686.html

https://www.tagesspiegel.de/berlin/14-millionen-dateien-aus-berliner-it-netz-erbeutet-hacker-veroffentlichen-geheime-daten-fur-den-verteidigungsfall-als-download-16015921.html

https://www.golem.de/news/berliner-verwaltungsdaten-geleakt-der-daten-super-gau-aus-der-hauptstadt-2609-212694.html 

https://x.com/PeterTiede/status/2096663305196798131

https://www.rbb24.de/politik/beitrag/2026/09/berlin-hackerangriff-schutz-software-streit-senat-bezirke.html

https://www.golem.de/news/hackerangriff-in-berlin-bezirk-verweigert-einsatz-von-crowdstrike-2609-212690.html

https://www.euronews.com/next/2026/09/05/berlin-cyberattack-hackers-leak-highly-sensitive-data-across-dark-web

17:28 KI und Produktivitätszuwächse in Unternehmen

https://lazy-fox.ai/blog/why-ai-transformations-fail.html

02:27 – Der große Meta-Vergleich: 18 Milliarden statt einer Billion

• AP über KOB 4: Meta reaches landmark $18 billion settlement with states in trial over teen social media addiction
• Free Speech Center: Meta reaches landmark $18 billion settlement

• AP über KOB 4: Details des Vergleichs

• Texarkana Gazette (AP): Meta reaches landmark $18 billion settlement

17:49 – Nvidia kauft Hugging Face für 13 Milliarden

• CNBC: Nvidia agrees to buy Hugging Face for $12.9 billion, report says
• TechCrunch: Nvidia closes in on Hugging Face acquisition
• Forbes: Nvidia Has Reportedly Agreed To Buy Hugging Face For $13 Billion

19:46 – Nvidia investiert in Perplexity

• The Information: Nvidia Discusses Perplexity Investment at $30 Billion-Plus Valuation
• Reuters über Yahoo Finance: Nvidia discusses Perplexity investment at $30 billion-plus valuation

24:30 – WSJ: Nvidias Rechenzentrums-Deals und die Wettbewerbsfrage

• Techstrong.ai: NVIDIA Pauses $36 Billion Revenue-Sharing Financing Model Amid Antitrust Scrutiny
• TNW: Nvidia has paused parts of the revenue-sharing programme it launched in July
• Tom's Hardware: Nvidia denies pausing AI cloud commitments initiative after reported partner backlash

27:36 – Anthropic: 30 Billionen TAM und die MHS-Ankündigung — mit Agnieszka

• Anthropic: Previewing the Model Hardware Standard
• CNBC: Anthropic pushes into physical world with new standard to help AI agents operate machines
• Fortune: Anthropic makes first move into physical AI with universal standard
• The Register: Anthropic proposes plumbing spec to link AI agents to lab kit and robots
• Fortune: Anthropic's $30 trillion market size estimate is outlandish. That may be the point.
• Quartz: Anthropic pitches IPO investors on $30 trillion market opportunity

38:23 – Nvidia-Quartalszahlen: der Ausbruch aus dem Negativstrudel

• CNBC: Nvidia earnings takeaways – Huang forecasts 70% fiscal 2028 revenue growth, far above estimates
• Kiplinger: Nvidia Earnings – Updates and Commentary August 2026

• Motley Fool: Nvidia (NVDA) Q2 2027 Earnings Call Transcript

• CNBC: Huang forecasts 70% fiscal 2028 revenue growth

47:17 – CrowdStrike: Cybersecurity als Boom-Markt

• StockStory über FinancialContent: CrowdStrike Surprises With Q2 CY2026 Sales, Stock Jumps 11.7%
• Timothy Sykes: CrowdStrike Stock Jumps As AI Security Fuels Record Growth

49:44 – SaaS-Apokalypse? Salesforce-Zahlen und Agentforce

• CNBC: Salesforce stock jumps 18% on AI growth and Anthropic investment gain
• CNBC: Marc Benioff is getting his mojo back as Salesforce's AI strength quiets skeptics
• TechTimes: Salesforce Launches Claudeforce, Posts Four-Year Revenue High as AI Threat Thesis Collapses

• Motley Fool: Here's Why Salesforce Stock Soared Today

00:00 Intro

00:16 Themenüberblick der Woche

04:26 Moderna & Merck: Personalisierte Krebsimpfung besteht Phase 3

• Dermatology Times: Personalized mRNA-Based Melanoma Vaccine Meets Primary Endpoints

13:49 Age of Always On: Meta Ray-Ban und der Kipppunkt

• EFF/State of Surveillance: Meta Shipped Facial Recognition Code to Millions of Ray-Bans
• Next Reality: Meta Ray-Ban Glasses Facial Recognition Quietly Removed After Outcry

17:24 ICE verbietet Mitarbeitern die Meta-Brille

• Engadget: ICE agents can't wear Meta glasses while they work, official memo warns⁠
• ⁠Gizmodo: Even ICE Thinks Smart Glasses Are a Privacy Liability⁠

19:49 Apple: AirPods mit Kamera

• MacRumors: Apple's Camera-Equipped AirPods Confirmed – See Them in Action⁠
• ⁠TechCrunch: Apple's new macOS update reportedly contains a video of AirPods with a camera⁠

23:15 Voice-Recorder als eigene Hardware-Kategorie: Plaud, Pocket & Co.

• ⁠TechCrunch: Pocket raises $11M in bet on rising demand for AI note-taking devices⁠
• ⁠AI Chat Daily: Plaud crosses $100M ARR after shipping 2M AI notetakers⁠

27:36 Heimlich mitschneiden: Recht, Ethik und Note-Taker

• TechTimes: AI Voice Recorder Raises $11M – Using It Without Consent Is a Felony in These States

31:41 Always On im Unternehmen: Real-Time-Steuerung und Second Order Effects

38:30 OpenAI zieht die Handbremse: Trainingsstopp und Modell Astra

• OpenAI: Pacing model development in an era of cyber-critical capabilities
• Axios: OpenAI Astra may have hit critical cyber threshold, prompting safety overhaul

• Forbes: OpenAI Paused AI Training For Two Weeks – Here's What That Means

41:46 Alibaba Qwen 3.8: Frontier-Niveau als Open Source

• MarkTechPost: Alibaba Qwen Releases Qwen3.8-Max
• Open Source For You: Alibaba Debuts Frontier-Scale Open Source Qwen3.8-Max

43:12 Cybersecurity: Wenn Patch Day zu Exploit Day wird

• Plain English Podcast mit Alex Stamos
• CyberScoop: Security leaders say the next two years are going to be 'insane' (Mandia, Stamos, Adamski)

• SecurityWeek: August 2026 Patch Tuesday – Microsoft Fixes 421 CVEs, One Exploited Zero-Day

48:40 OpenAI: Datenschutz für Geschäftskunden

• Help Net Security: OpenAI previews privacy-focused system for detecting AI misuse
• Digital Trends: OpenAI wants to monitor AI abuse without forcing customers to hand over their data

52:05 ChatGPT "Computer History": Keylogger als Feature

• The Decoder: OpenAI's Computer History turns your clicks and keystrokes into a searchable ChatGPT memory timeline
• TNW: OpenAI's new ChatGPT feature logs your keystrokes and stores them in plain text

54:56 OpenAI: Börsengang 2027, Zahlen und Werbung

• CNBC: OpenAI 'will be a public company in 2027' or sooner, CFO Friar tells employees

• Technology.org: OpenAI Cuts GPT-5.6 Sol Developer Pricing by More Than 20%
• Neowin: OpenAI expands ChatGPT ads to 31 European markets

58:35 ChatGPT for Teens und Elternkontrolle

• OpenAI: Introducing ChatGPT for Teens
• OpenAI: Introducing parental controls
• gHacks: OpenAI Launches ChatGPT for Teens With Study-Focused Features

1:02:52 Google kauft die Daten der bankrotten Spirit Airlines

• Axios: Google buys Spirit Airlines emails, chats, documents out of bankruptcy

• Forbes: Google Paying $10 Million For Spirit Airlines' Data To Train AI

1:05:03 Daten als Gold: Mercor, Simple Closure & Monetarisierung

• Fast Company: Shuttered startups are selling old Slack chats and emails to AI companies

1:10:39 Digitale Zwillinge: Harvard klont Dozenten

• TechCrunch: Harvard's $699 startup bootcamp offers AI avatars of its instructors

1:13:18 Pew Research: Wie viel des Internets schreibt die KI?

• Pew Research Center: How Much of the Internet Is Written With AI?

• TechCrunch: A third of web pages published since ChatGPT launched were written by AI

1:18:37 Buchempfehlung: "Don't Be Evil"

1:23:57 Buchverkäufe im KI-Zeitalter – und Amazons Buchschredder

• tim.blog: Has AI Already Killed How-To Nonfiction? Sales Trends, My Personal Data, and What It Might Mean

• 404 Media: We Tracked a Shipment of Rare Books. It Ended at an Amazon AI Training Facility

00:15 Einführung

04:15 Unsichtbare Tinte für KI-Text: Anthropic watermarkt Claudehttps://www.anthropic.com/news/claude-text-watermarkhttps://the-decoder.com/anthropic-announces-watermark-detection-api-that-will-let-third-parties-detect-claudes-ai-texts/https://www.pangram.com/blog/what-does-your-ai-detection-score-meanhttps://www.forbes.com/sites/jonmarkman/2026/08/13/anthropic-eyes-2-trillion-in-october-ipo-a-record-breaking-debut/

32:02 Suno watermarkt Songs, Spotify labelt KI-Künstlerhttps://techcrunch.com/2026/08/06/amid-legal-battles-suno-says-it-will-start-watermarking-songs/https://techcrunch.com/2026/08/11/spotify-will-label-ai-persona-profiles-and-exclude-their-music-from-recommendations/

38:44 Das Wort der Woche: Harness. Grok Bot und DeepSeek Harnesshttps://deepseek.com/harness/en/https://thenewstack.io/deepseek-harness-open-source-plugins/https://docs.x.ai/grok-bot/computer-and-appshttps://arxiv.org/abs/2606.10106

48:15 Neues open-weight Modell von Meta und ein Manifest von Mark Zuckerberghttps://www.meta.com/thefutureisforeveryone/https://www.axios.com/2026/08/10/zuckerberg-ai-manifesto-metahttps://nmdoj.gov/press-release/court-orders-meta-to-pay-942-million-and-overhaul-protections-for-children-on-facebook-and-instagram-in-landmark-new-mexico-ruling/

56:56 WorldClaw: Tencent baut 3D-Welten und benutzt Claude als Gehirnhttps://arxiv.org/abs/2608.05248https://tencent-hunyuan.github.io/Hunyuan3D-WorldClaw/

59:10 Die Neuausrichtung von Mistralhttps://www.digitalapplied.com/blog/mistral-hosts-rival-open-models-regional-endpoints-2026https://techcrunch.com/2026/03/30/mistral-ai-raises-830m-in-debt-to-set-up-a-data-center-near-paris/

01:01:58 Gebärdensprache auf dem Handy: Googles SL2Thttps://deepmind.google/blog/putting-sign-language-ai-into-users-hands/

01:03:09 Wo geht's hin mit Lovable, Workday und Canvahttps://techcrunch.com/2026/08/12/lovable-confirms-new-13-3b-valuation-raises-another-400m/https://www.axios.com/2026/08/14/silver-lake-workday-take-privatehttps://cryptobriefing.com/canva-cuts-revenue-forecast-ai-costs/

01:10:32 Truth APIhttps://www.cnn.com/2026/08/11/business/wall-street-trump-truth-socialhttps://fortune.com/2026/08/12/trump-media-truth-api-insider-trading/

01:14:22 Buchempfehlung: Der Mann, der seine Frau mit einem Hut verwechseltehttps://www.oliversacks.com/oliver-sacks-books/the-man-who-mistook-his-wife-for-a-hat/

00:15 – Intro: Die Themen der Woche im Überblick

03:29 – Google-Beben: Jeff Dean gründet Startup, Demis Hassabis übergibt das Tagesgeschäft

• Google's AI reshuffle: Chief scientist Jeff Dean exits and Demis Hassabis steps down as DeepMind CEO (CNBC)
• Jeff Dean and other top AI researchers are leaving Google to launch their own startup (TechCrunch)
• Demis Hassabis' new Google DeepMind role explained (CNBC)

06:31 – DeepMinds Anfänge & warum Google die Sprachmodell-Welle verpasste

• The Infinity Machine: Demis Hassabis, DeepMind, and the Quest for Superintelligence – Sebastian Mallaby (Penguin Random House)
• The Infinity Machine (Goodreads)

14:27 – Braucht Google überhaupt das führende KI-Modell? „Good enough" reicht vielleicht schon

16:36 – Der Exodus: Noam Shazeer zu OpenAI, John Jumper zu Anthropic

• Google Gemini co-lead Noam Shazeer leaves for OpenAI (CNBC)
• Star Google Researcher Joins OpenAI in Coup for ChatGPT Creator (Bloomberg)

• Nobel laureate John Jumper is leaving DeepMind for rival Anthropic (TechCrunch)

20:04 – Jeff Deans neues Startup „Discovery Loop": KI automatisiert die wissenschaftliche Methode

• Jeff Dean Leaves Google to Automate the Scientific Method With Discovery Loop (Unite.AI)
• Jeff Dean leaving Google after 27 years to co-found Discovery Loop (Quartz)

21:55 – Anthropic baut eigene Chips: Gespräche mit Samsung

• Anthropic in Talks With Samsung to Manufacture Custom AI Chip (The Information)

29:00 – Wettrüsten ums „gefährlichste" KI-Modell: OpenAI, Anthropic und jetzt auch Meta

• Déjà Vu? Meta's AI Escapes Testing Lab in Hacking Joyride (Dark Reading)
• Meta AI breaches external firm during security testing sandbox error (NPR)

32:46 – NVIDIAs „Open Secure AI Alliance": Verteidiger brauchen auch Frontier-KI

• Industry Leaders Join Open Secure AI Alliance for AI Safety and Security (NVIDIA Blog)
• NVIDIA Forms 37-Member Open Secure AI Alliance and Open-Sources NOOA Framework (The Hacker News)
• Nvidia, SpaceX, Microsoft launch AI safety initiative as OpenAI cyberattack fallout continues (CNBC)

34:09 – Palo-Alto-Networks-CEO Nikesh Arora: Sind die Modell-„Ausbrüche" nur PR?

• 20VC: Palo Alto Networks CEO Nikesh Arora – The Twenty Minute VC mit Harry Stebbings (Apple Podcasts)
• 20VC: Palo Alto Networks CEO Nikesh Arora on How to Create and Sustain Competitive Advantage and Defensibility (Spotify)

36:56 – Bending Spoons übernimmt Airtable für 1,3 Milliarden Dollar

• Bending Spoons makes first post-IPO acquisition with $1.3 billion Airtable deal (CNBC)
• Bending Spoons to buy Airtable for $1.28B (TechCrunch)

45:47 – Shopify: KI-Traffic und -Bestellungen verdreifachen sich

• Shopify Q2 2026 slides: AI commerce drives 3x traffic surge (Investing.com)
• Shopify's AI Traffic Triples as Shoppers Skip the Search Bar (PYMNTS)
• Shopify Credits AI for 34% Revenue Growth in Q2 2026 (Retail TouchPoints)

49:47 – Atlassian: Quartalszahlen lassen Aktie um 35 % steigen

• Here's Why Atlassian Stock Surged 35% Last Friday (TIKR)
• Atlassian Earnings Call Highlights Cloud and AI Surge (The Globe and Mail)

57:46 – Notion im KI-Zeitalter: Bleibt das „Second Brain" relevant?

1:04:33 – MySpace-Comeback: Ein Social Network ohne Algorithmus

• MySpace confirms it is coming back, without Myspace Tom (We Rave You)
• Myspace eyes comeback as an 'antidote' to social media fatigue — but it may be a long shot (CNBC)
• Myspace wants to sell you a feed without an algorithm. Brussels is about to make that mandatory anyway (The Next Web)

1:12:50 – Buchempfehlung: „The Infinity Machine" von Sebastian Mallaby

• The Infinity Machine: Demis Hassabis, DeepMind, and the Quest for Superintelligence (Penguin Random House)
• sebastian mallabys new cfr book details the rise of deepminds demis hassabis (Council on Foreign Relations)

00:01 – Intro: Die Themen der Woche im Überblick

03:07 – Die AI-Wette in der Crunchtime

05:25 – NVIDIAs 250-Milliarden-Dollar-Garantie für OpenAIs Rechenzentren

• Nvidia in talks with OpenAI to guarantee $250 billion financing for data center, WSJ reports (Yahoo Finance)
• Big Tech's AI Debt Binge Tests High-Grade Market, Barclays Says (Yahoo Finance)

06:50 – Microsoft-Zahlen: Azure wächst 43 %, größter Tagesgewinn der Geschichte

• Microsoft's stock has biggest one-day gain since 2008, adding $480 billion in market value (Fortune)
• Microsoft (MSFT) Q4 earnings report 2026 (CNBC)

08:24 – Meta vs. Microsoft: Warum die Börse so unterschiedlich reagiert

• Meta's stock drops on disappointing guidance, dwindling free cash flow (CNBC)

12:56 – Microsofts Klumpenrisiko: 625 Milliarden Dollar Auftragsbestand, 45 % davon OpenAI

• Microsoft (MSFT) Q4 earnings report 2026 (CNBC)

14:15 – Amazon-Zahlen: AWS wächst 37 %, Capex steigt auf 220 Milliarden Dollar

• Amazon hikes 2026 capex to $220 billion due to higher memory costs (CNBC)
• AWS Backlog Hits $496B as Amazon Raises AI Spend to $220B and Capacity Runs Short (Tech Times)

15:36 – Leopold Aschenbrenner: Vom KI-Wunderkind zum Margin Call

• How Leopold Aschenbrenner built a $45 billion AI hedge fund — and lost most of it in days (CNBC)
• AI investor Leopold Aschenbrenner forced to unwind all public stock positions after steep losses, sources say (CNBC)
• Timing - and task number one... https://x.com/dnapway/status/2082876150322540657 https://x.com/tbpn/status/2083226453509030285

27:30 – Larry Ellison und der historische Oracle-Crash

• Oracle's stock crash looks embarrassing for billionaire founder Larry Ellison (Yahoo Finance)

30:02 – Die zentrale Frage: Data-Center-Boom vs. Enterprise-Returnhttps://x.com/raph_dornano/status/2082164433548443726

32:31 – OpenAI Presence: Der Vorstoß in die Unternehmen

• Introducing OpenAI Presence (OpenAI)

• OpenAI tries the consulting path with 'Presence', charging enterprises boots-on-the-ground prices to deploy agents (The Register)

35:15 – Wer ist gefährlicher? Anthropic und OpenAI melden Modell-Ausbrüche

• Anthropic's Claude AI Broke Into Three Companies During Security Tests (Forbes)
• After OpenAI disclosure, Anthropic says Claude also hacked outside systems (Al Jazeera)

38:25 – "Prompt Injection als Role Confusion": Warum Sprachmodelle sich nie ganz absichern lassen

• Prompt Injection as Role Confusion (arXiv)
• Interesting Paper Exploring Prompt Injection (Schneier on Security)
• Security researchers tricked LLMs into giving them cocaine recipes by abusing role models for prompt injection (The Register)

44:26 – Satya Nadellas Datenparadox: Was ist das IP eines Unternehmens in der AI-Ära?https://lazy-fox.ai/blog/nadella-reverse-paradox.html

46:32 – EU AI Act: Kennzeichnungspflicht für KI-Inhalte seit 2. August

• The EU AI Act's Transparency Rules: A Practical Guide to Article 50 (artificialintelligenceact.eu)
• EU AI Labeling Rules Start August 2: What Article 50 Requires (progressiverobot.com)

52:45 – LinkedIns neuer "Seems like AI Slop"-Button

• LinkedIn adds a button to report AI-generated 'slop' (TechCrunch)
• LinkedIn adds a 'seems like AI slop' button after blocking billions of automated comment attempts (Fortune)

54:30 – Der geplatzte Buchdeal und die Frage: Was ist eigentlich "Slop"?

• Seven-Figure Book Deal Cancelled Over AI Suspicions Raises Questions About AI Guardrails (Publishers Lunch)

1:03:28 – Jack Dorseys "Buzz": Ein KI-nativer Slack-Herausforderer

• Jack Dorsey is taking on Slack with Buzz, a group chat platform for teams and their AI agents (TechCrunch)

1:08:35 – Google stellt Gemini Robotics ER 2 vor

• Introducing Gemini Robotics ER 2 (Google Blog)

• Gemini Robotics 2 Expands Google's AI Capabilities for Humanoid Robots (Bloomberg)

1:10:25 – Travis Kalanick: 1,7 Milliarden Dollar für Robotik-Startup Atoms

• Travis Kalanick's robotics company raises $1.7B, led by a16z (TechCrunch)

1:11:41 – Buchempfehlung: "Der Gang vor die Hunde" von Erich Kästner

00:15 Einführung

04:27 OpenAIs Model haut aus dem Testkäfig ab und hackt Hugging Face

https://www.theguardian.com/technology/2026/jul/22/openai-says-its-models-went-rogue-and-hacked-startup-in-unprecedented-incidenthttps://huggingface.co/blog/security-incident-july-2026https://openai.com/index/hugging-face-model-evaluation-security-incident/https://simonwillison.net/2026/Jul/22/openai-cyberattack/https://time.com/article/2026/07/24/openai-hugging-face-attack/https://epoch.ai/gradient-updates/openai-accidentally-hacked-hugging-face

15:45 Sanktionen gegen Chinesische LLMs und die Frage der Legalität der Destillation

https://techcrunch.com/2026/07/22/treasury-threatens-sanctions-after-white-house-claims-moonshot-distilled-anthropics-fable/https://techcrunch.com/2026/07/23/experts-say-exploiting-anthropics-fable-isnt-how-kimi-k3-got-so-good/https://www.microsoft.com/en-us/corporate-responsibility/topics/open-weight/https://www.nytimes.com/2026/07/25/technology/open-source-silicon-valley-china.htmlhttps://simonwillison.net/2026/Jul/16/kimi-k3/

25:24 Google Quartalszahlen und neue, günstigere Modelle

https://9to5google.com/2026/07/22/alphabet-q2-2026-earnings/https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-6-flash-3-5-flash-lite-3-5-flash-cyber/https://techcrunch.com/2026/07/21/google-releases-three-new-gemini-models-but-no-3-5-pro/https://artificialanalysis.ai/models/gemini-3-6-flash

34:14 Anthropic: Opus 5 schlägt Fable in fast allen Benchmarks

https://www.anthropic.com/news/claude-opus-5https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-opus-5https://simonwillison.net/2026/Jul/24/introducing-claude-opus-5/https://support.claude.com/en/articles/12512198-how-to-create-custom-skillshttps://www.buseybench.com/

48:11 Kleine Firmen und KI: nicht Entlassungswelle, sondern Notwehr

https://www.theguardian.com/business/2026/jul/26/small-businesses-aihttps://ec.europa.eu/eurostat/statistics-explained/index.php?title=Use_of_artificial_intelligence_in_enterpriseshttps://www.bitkom.org/Presse/Presseinformation/Digitalisierung-der-Wirtschaft-Unternehmen-beschaeftigen-sich-mit-KIhttps://www.ifo.de/pressemitteilung/2026-06-16/selbstaendige-setzen-staerker-auf-kuenstliche-intelligenzhttps://papers.ssrn.com/sol3/papers.cfm?abstract_id=5240924

56:00 About You schließt im Oktober das Fotostudio und generiert Produktbilder mit KI

https://www.ibusiness.de/aktuell/db/653402sh.htmlhttps://www.scayle.com/studios/

58:24 Warum sich Prozessanwälte gegen die selbstfahrenden Fahrzeuge währen

https://marginalrevolution.com/marginalrevolution/2026/07/trial-lawyers-lobby-against-autonomous-vehicles.htmlhttps://waymo.com/safety/impact/

01:02:11 Buch der Folge: Inside the Box — David Epstein

https://davidepstein.com/inside-the-box/https://www.penguinrandomhouse.com/books/737494/inside-the-box-by-david-epstein/

00:15 Einführung

04:41 Volatilität an den Märkten und deren Auswirkungenhttps://www.nytimes.com/2026/07/17/business/stocks-ai-tech-wall-street.htmlhttps://thehill.com/policy/technology/5624075-ai-bubble-market-volatility/

12:27 China holt massiv auf: Kimi K3 nur knapp hinter Fable und GPT 5.6https://techcrunch.com/2026/07/16/moonshots-upcoming-kimi-3-is-expected-to-close-the-gap-with-anthropics-opus-4-8/https://www.cnbc.com/2026/07/17/moonshot-ai-kimi-k3-model-openai-anthropic-china.htmlhttps://simonwillison.net/2026/Jul/16/kimi-k3/https://the-decoder.com/kimis-open-model-k3-nears-gpt-5-6-sol-and-fable-5-while-signaling-the-end-of-super-cheap-chinese-ai/

https://www.linkedin.com/feed/update/urn:li:activity:7485227891765665792/

19:47 Inkling: ein Modell, das absichtlich nicht das beste isthttps://thinkingmachines.ai/news/introducing-inkling/https://techcrunch.com/2026/07/15/thinking-machines-amps-up-its-bet-against-one-size-fits-all-ai-with-its-first-open-model-inkling/https://artificialanalysis.ai/articles/thinking-machines-has-released-inkling-the-new-leading-u-s-open-weights-model

25:28 Die Compute-Rechnung: wer gewinnt langfristig?https://www.nytimes.com/2026/07/17/technology/meta-anthropic-ai-computing-power.htmlhttps://www.cnbc.com/2026/07/17/anthropic-meta-ai-compute.htmlhttps://www.bloomberg.com/news/articles/2026-05-20/anthropic-to-pay-spacex-nearly-45-billion-for-computing-dealhttps://www.axios.com/2026/05/14/anthropic-claude-price-openai-tokens

31:12 Loop engineering: wie machen wir das bei mikahttps://tech-blog.getmika.de/posts/notes-from-an-autonomous-build/

43:10 Grok kopiert ganze Entwickler-Repositorieshttps://cybernews.com/ai-news/grok-build-git-repository-upload/https://thehackernews.com/2026/07/grok-build-uploads-entire-git.htmlhttps://thenextweb.com/news/grok-build-uploaded-entire-git-repositories-secrets

48:14 Claude: integrierter Browser und 1Passwordhttps://code.claude.com/docs/en/whats-newhttps://1password.com/blog/1password-for-claudehttps://techcrunch.com/2026/07/14/openais-first-hardware-device-is-reportedly-a-screenless-speaker-that-can-move/

56:10 Wohin entwickelt sich die Google KI?https://techcrunch.com/2026/07/16/googles-ai-mode-now-lets-you-link-and-interact-with-select-apps/https://techcrunch.com/2026/07/16/google-continues-its-renaming-streak-by-turning-notebooklm-to-gemini-notebook/https://9to5google.com/2026/06/30/notebooklm-short-video-overviews/https://techcrunch.com/2026/07/16/google-vids-now-lets-you-star-in-your-own-ai-videos/https://www.axios.com/2026/07/10/meta-ai-image-consenthttps://theaiinsider.tech/2026/07/13/meta-removes-controversial-instagram-ai-photo-feature-following-widespread-backlash/

01:00:34 AGI, dysthopische Claude-Werbung und institutioneller Handlungsbedarfhttps://digitaleconomy.stanford.edu/news/wemustactnow/https://demishassabis.substack.com/p/a-framework-for-frontier-ai-and-the-dawning-of-a-new-agehttps://time.com/article/2026/07/16/anthropic-ai-marketing-advertisement/https://techcrunch.com/2026/07/14/anthropics-newest-ad-is-creeping-people-out/

01:08:32 Prominente Übernahmepläne von Stripe und Uberhttps://techcrunch.com/2026/07/15/stripe-and-advent-reportedly-offered-to-buy-paypal-for-around-53-4b/https://www.sec.gov/Archives/edgar/data/0001543151/000155278126000382/e26302_ex99-1.htm

01:11:57 Buchbesprechung: After Nationshttps://www.penguinrandomhouse.com/books/536696/after-nations-by-rana-dasgupta/https://www.newstatesman.com/culture/books/2026/03/in-time-of-the-breaking-of-nation-states