Hedge Funds reduzieren massiv Tech-Exposure – laut Goldman Sachs die größte 2-Wochen-Reduktion seit einem Jahrzehnt. Nur Profit-Taking nach der Rally oder Beginn einer Rotation? #Tech #Aktien #Finanzmärkte #HedgeFunds #Börse
Hedge funds are rushing to reduce tech exposure:
Hedge funds just posted their largest 2-week reduction in US information technology exposure over the last decade, excluding the meme stock frenzy in early 2021.
This was driven by long sales outpacing short covers at a ratio of 1.5 to 1.
Nearly every subsector saw exposure cuts, led by Semiconductors and Semi Equipment via long sales.
Furthermore, hedge funds sold Magnificent 7 stocks in 4 of the last 5 trading sessions.
Hedge funds are cashing-in massive profits in tech.
42
AI ist nicht mehr nur Software – wir stehen vor dem größten Shift seit der frühen Cloud-Ära. OpenAI zeigt: AI wird zunehmend zu einem Kapital-, Energie- und Chip-Game. 💡
#AI #TechInvesting
1
17
Die frühe Cloud-Ära erinnerte uns: massive Investitionen in Infrastruktur, Monetarisierung oft erst Jahre später. OpenAI zeigt dasselbe Muster: hohe Capex, langsame Einnahmen, strategische Geduld gefragt. 🏗️ #AI #TechInvesting
1
13
Strukturell gewinnen Hyperscaler, Chip-Oligopole und Energie-Player. Reine Model-Bauer werden zu großen Abnehmern.
AI skaliert nicht wie Software, sondern wie neue industrielle Infrastruktur. ⚡
Wer sieht das auch so? #TechStrategy
15
Markus retweeted
OpenAI and Anthropic engineers don't prompt like everyone else.
I've been reverse-engineering their techniques for 2.5 years across all AI models.
Here are 5 prompting methods that get you AI engineer-level results:
Markus retweeted
𝗦𝗵𝗼𝗽𝗶𝗳𝘆 𝗵𝗶𝘁 𝟰𝟴𝟵𝗠 𝗿𝗲𝗾𝘂𝗲𝘀𝘁𝘀/𝗺𝗶𝗻 𝗱𝘂𝗿𝗶𝗻𝗴 𝗕𝗹𝗮𝗰𝗸 𝗙𝗿𝗶𝗱𝗮𝘆 𝗼𝗻 𝗮 𝗺𝗼𝗻𝗼𝗹𝗶𝘁𝗵
Many engineers assume this scale forces microservices.
Shopify keeps proving a different playbook: 𝗼𝗻𝗲 𝗱𝗲𝗽𝗹𝗼𝘆, 𝘀𝘁𝗿𝗼𝗻𝗴 𝗯𝗼𝘂𝗻𝗱𝗮𝗿𝗶𝗲𝘀, 𝗵𝘂𝗴𝗲 𝗶𝗻𝗳𝗿𝗮
𝗕𝗹𝗮𝗰𝗸 𝗙𝗿𝗶𝗱𝗮𝘆 𝟮𝟬𝟮𝟱, 𝗯𝘆 𝘁𝗵𝗲 𝗻𝘂𝗺𝗯𝗲𝗿𝘀:
🔹 $14.6B merchant sales (+27% YoY)
🔹 Edge: 489M requests/min peak (312M avg across BFCM)
🔹 App servers: 117M requests/min peak (+40% YoY)
🔹 API: 31.8M requests/min peak
🔹 Observability: 11TB logs/min at peak
🔹 Kubernetes: 3.18M CPU cores at peak
🔹 DB (MySQL 8): 53.8M queries/sec, plus 4.28B row ops/sec at peak
🔹 Writes: 2M writes/sec at peak
🔹 CDN: 183M requests/min, 97.8% served from cache
🔹 Async: 23.2M jobs/min at peak
🔹 Streaming: Kafka + Flink, 150 MB/sec processed; analytics latency improved 103×
Why a modular monolith works here:
➡️ One codebase, one release train, one primary runtime.
➡️ Clear domain components (orders, inventory, billing…) with enforced interfaces.
➡️ Rails Engines + tooling (Packwerk) to keep boundaries real, not “handshake agreements”.
👉 𝗧𝗮𝗸𝗲𝗮𝘄𝗮𝘆: scaling often fails from 𝗯𝗼𝘂𝗻𝗱𝗮𝗿𝘆 𝗱𝗲𝗯𝘁, not “monolith vs microservices.”
Pick the simplest architecture that lets teams ship fast, then enforce boundaries hard.
#softwareengineering #technology #programming
Markus retweeted
Last month, Google dropped something interesting:
five AI Agent papers released across five consecutive days, one per day, each digging into a different part of how agents should be built, evaluated, secured, and deployed.
No big splash, just a steady rollout of more than 250 pages of highly technical material.
Here’s the distilled version of what those five drops covered:
Read. Learn. Bookmark.
Markus retweeted
It's more like 60/40
Coding agents could never work as workflows
Lots of workflows I run would be much worse as agents
But it's true that most people underestimate workflows
Markus retweeted
After an interview with @karpathy, everyone is talking about what AI agents can/can't do.
But an opinion without data is just a hypothesis.
So, I tested 3x185 workflow executions for a market researcher agent.
The results have shocked me🧵
Markus retweeted
kvcached works directly with vLLM and you can use it to serve multiple models on the same GPU. They will share unused KV cache blocks. Check it out!
🚀 End the GPU Cost Crisis Today!!!
Headache with LLMs lock a whole GPU but leave capacity idle? Frustrated by your cluster's low utilization?
We launch kvcached, the first library for elastic GPU sharing across LLMs.
🔗 github.com/ovg-project/kvcac…
🧵👇 Why it matters:
🚨 Breaking: Amazon plant bis zum Jahr 2033 unglaubliche 600.000 Jobs in den USA durch Roboter zu ersetzen.
Details zu Amazons Automatisierungsplänen: golem.de/news/roboter-statt-…
Was denkst du? #Automatisierung #Amazon
Image via @grok / @xai
1
23
Markus retweeted
🚀 DeepSeek-OCR — the new frontier of OCR from @deepseek_ai , exploring optical context compression for LLMs, is running blazingly fast on vLLM ⚡ (~2500 tokens/s on A100-40G) — powered by vllm==0.8.5 for day-0 model support.
🧠 Compresses visual contexts up to 20× while keeping 97% OCR accuracy at <10×.
📄 Outperforms GOT-OCR2.0 & MinerU2.0 on OmniDocBench using fewer vision tokens.
🤝 The vLLM team is working with DeepSeek to bring official DeepSeek-OCR support into the next vLLM release — making multimodal inference even faster and easier to scale.
🔗 github.com/deepseek-ai/DeepS…
#vLLM #DeepSeek #OCR #LLM #VisionAI #DeepLearning
Markus retweeted
New York 🇺🇸 Such an office room
Readers added context they thought people might want to know
This picture is AI generated
wikipedia.org/wiki/Generativ…
Reliability Engineering: So important. Especially today. 🚀
#ReliabilityFirst #CloudComputing
4
Markus retweeted
1) Amazon Aktie fällt 8% 📉 Die Gründe:
🛜AWS enttäuscht: Wachstum ist mit 17,5 % ist langsamer als bei MSFT und GOOG.
💸 Hohe Ausgaben: Amazon plant über 100 Mrd. USD Investitionen
⚠️ Verhaltene Prognose: Deshalb liegt der operative Gewinn für Q3 lag unter den Markterwartungen
🧠 Unklare KI-Strategie: Investitionen in KI steigen, aber Monetarisierung ist vage
🌍 Zoll- und Konjunkturrisiken: Neue US-Zölle und globale Unsicherheiten
📊 Marktumfeld negativ: Tech-Sektor heute unter Druck






























