Inlägg

Visar inlägg med etiketten Fang

Cevians vd och medgrundare Christer Gardell vill inte dödförklara AI-hajpen, men ser Svart Svan

Bild
   ”AI kommer att ha stor inverkan på allt möjligt men sektorn är galet övervärderad. Sedan är det mycket spekulativa flöden och kasinopengar – så fort något nytt kommer in på börsen vill många vara med och spela – så där kommer det att smälla ordentligt”, säger han. Hur sårbar är du själv för en AI-smäll? ”Inte alls, vi håller inte på med sådant där.” Christer Gardell är mer orolig för västvärldens relation med Kina framöver. ”Skulle väst gå för hårt åt och isolera Kina för mycket riskerar världen att slitas isär i ett ‘vi och dom’.  Det är den verkligt stora svarta svanen. Dagens industri 14 februari 2025 https://www.di.se/digital/gardell-varnar-dar-kommer-det-smalla-ordentligt/ Tillbaka till Rolfs länktips 14 Februari 2025 https://englundmacro.blogspot.com/2025/02/rolfs-lanktips-14-februari-2025.html

By 2028, US will need to boost annual electricity generation by as much as 26 per cent

Bild
  from 2023 levels to meet demand, far higher than any supply growth the US has achieved in the past two decades Nearly 45 per cent of this demand growth is being driven by the construction of power-hungry data centres required to support the rollout of AI technologies.     Financial Times 4 February 2025 https://www.ft.com/content/47dbfee3-5517-43c6-85ab-dfe86d2d4085 Tillbaka till Rolfs länktips 4 Februari 2025 https://englundmacro.blogspot.com/2025/02/rolfs-lanktips-4-februari-2025.html

Den nya EU-backade språkmodellen ”Open Euro”

 ”Jag har propagerat länge för att det behövs en stor satsning, inte tio små”, säger professorn och Silo-grundaren Peter Sarlin som ska leda EU-forskarna. EU-kommissionens satsning ”Euro HPC” – kort för ”high performance computing”. Donald Trump presenterade nyligen projektet Stargate – ett samarbete mellan investerare, AI-experter och molnjättar som tillsammans säger sig villiga att lägga över 500 miljarder dollar över fyra år på amerikansk AI-utveckling, inte minst genom att bygga datacenter och infrastruktur. EU-kommissionens initiala bidrag till Open Euro är 54 miljoner euro, motsvarande 620 miljoner kronor. Det inkluderar dock inte kostnaderna för att träna de kommande AI-modellerna. ”Det är huvudsakligen för löner”, förklarar Peter Sarlin Peter Sarlin blir i egenskap av professor och grundare av finländska Silo – som jobbar nära Wallenbergs Combient-nätverk DI 3 februari 2025 https://www.di.se/digital/de-bygger-europas-deepseek-ai-profilen-ska-leda-forskarna/ EuroHPC JU is a ...

Bill Gates född 1955 : I dag hade jag fått diagnosen autism

Bild
  DN 3 februari 2025 https://www.dn.se/varlden/bill-gates-i-dag-hade-jag-fatt-diagnosen-autism/

Alibaba’s Qwen team released a new family of AI models, Qwen2.5-VL

Bild
...  that can perform a number of text and image analysis tasks. The models can parse files, understand videos, and count objects in images, as well as control a PC — similar to the model powering OpenAI’s recently launched Operator. Per the Qwen team’s benchmarking, the best Qwen2.5-VL model beats OpenAI’s GPT-4o, Anthropic’s Claude 3.5 Sonnet, and Google’s Gemini 2.0 Flash  on a range of video understanding, math, document analysis, and question-answering evaluations. https://techcrunch.com/2025/01/27/alibabas-qwen-team-releases-ai-models-that-can-control-pcs-and-phones/ DeepSeek Just Lost to Alibaba’s New AI! https://www.youtube.com/watch?v=0GQe7FUkZBA&t Tillbaka till Rolfs länktips 1-2 Februari 2025 https://englundmacro.blogspot.com/2025/02/rolfs-lanktips-1-2-februari-2025.html

“It took Deepseek about 12 seconds.”

Bild
  The blogger who helped spark Nvidia’s $600 billion stock collapse and a panic in Silicon Valley  Jeffrey Emanuel says Wall Street banks that are bullish on Nvidia ‘have absolutely no idea what they’re talking about’ Last Friday afternoon, Jeffrey Emanuel sat down in his Brooklyn apartment and started writing a blog post. For hours, he pounded away on his keyboard while his wife kept their young children occupied and brought him food.  Emanuel worked late into the night, and by early Saturday morning he had written nearly 12,000 words. Emanuel’s manifesto made the case for shorting the hottest company in the stock market. MarketWatch 1 February 2025 https://www.marketwatch.com/story/the-blogger-who-helped-spark-nvidias-600-billion-stock-collapse-and-a-panic-in-silicon-valley-52aba340 Whatever AI’s “Sputnik moment” ultimately means for stocks, the boost to labor productivity may come sooner than expected. In a post on Substack on Monday, Stony Brook University Professor o...

John Mauldin Why DeepSeek Is Bullish for the World

“There are decades when nothing happens, and weeks when decades happen,” says a quote from Vladimir Lenin, who may have copied it from someone else.  US companies like Nvidia (NVDA) would profit from making those chips, which other US companies would buy and use to develop AI applications in vast new US data centers.  The main worry was getting enough electricity to power it all. China, it was thought, would be hobbled for lack of the necessary chips, and thus present little competition. DeepSeek just poked a massive hole in that happy narrative.  When the topic involves China and technology, one of my go-to sources is Gavekal Research.  They know how the Chinese economy works and are really good at explaining it in ways Westerners can understand. Louis Gave knew DeepSeek was a gamechanger as soon as he saw it. He posted a report calling this Another Sputnik Moment. Gavekal’s Beijing-based tech analyst Tilly Zhang followed up with a deeper explanation. This is a good...

With DeepSeek, China innovates and the US imitates

Bild
Money doesn’t solve all problems, but $37.4 billion should solve a few.  That was what Microsoft and Meta have in combined capital spending in the December-ended quarter.  The expenditures were higher than what Wall Street had anticipated, not to mention nearly double from the same period the previous year.  The bulk of it is going to chips and data centers to power generative artificial intelligence services.  Wall Street Journal 29 January 2025 https://www.wsj.com/tech/ai/microsoft-and-metas-ai-splurge-still-not-enough-for-nvidia-916daca2 “You can see the computer age everywhere but in the productivity statistics,” wrote Nobel Prize-winning economist Robert Solow in 1987 DeepSeek has made us wonder, when does it happen for AI? AI Has Rocked the Stock Market, But What Will It Do for the Economy? Weekend Essay | The Big Take Bloomberg 31 January 2025 https://www.bloomberg.com/news/articles/2025-01-31/will-ai-take-our-jobs-3-scenarios-for-how-it-could-impact-the-econo...

Tech megacaps have sky-high multiples; can they really justify them?

Bild
  The asteroid hit of DeepSeek’s shoestring-budget artificial intelligence chatbot, which has inflicted massive damage on some US tech stocks, raised the stakes even further.  The tech megacaps have sky-high multiples; that’s nothing new. Can they really justify them? The market’s initial judgment is that they can — but that seems questionable.  Microsoft  Profits from such huge investments will not be instant, but the expectations baked into the company’s valuations leave little room for disappointment It’s true that you buy a company’s future earnings when you buy its stock. But it’s amazing that so much can be taken on trust. Somehow, even after the almighty DeepSeek shock, investors still seem prepared to give Big Tech the benefit of the doubt. John Authers Bloomberg 30 January 2025 https://www.bloomberg.com/opinion/articles/2025-01-30/ai-euphoria-saves-big-tech-from-so-so-earnings Tillbaka till Rolfs länktips 30 januari 2025 https://englundmacro.blogspot.com/202...

DeepThoughts on DeepSeek - 42

Bild
  In The Hitchhiker’s Guide to the Galaxy, humans build the greatest computer ever, called Deep Thought, to answer the “ultimate question of life, the universe and everything.”  After millions of years to think about it, it spits out the answer: 42.  It will take time to sort out how much DeepSeek changes the economics of AI. For now, here are some of the deeper thoughts as people try to work out the question they most need to ask. Markets bounced, as might be expected, without regaining all the ground lost on Monday.  The ongoing damage seems worse for the utility companies that stood to provide the extra power to keep AI data centers running.  If the Magnificents had formed a bubble, then, it still hasn’t burst.  But Monday’s sell-off shows that the story of Magnificent exceptionalism isn’t strongly rooted.  Diagnosing bubbles in real time is difficult. However, a market this dependent on a few stocks is a bad sign. John Authers Bloomberg 29 January ...

DeepSeek and Everyone’s Assumptions About AI Costs

Bild
  The Chinese upstart says its rival to ChatGPT comes at a fraction of the cost, raising questions about the rationale for stratospheric AI budgets  “We’re probably a year-plus ahead in models,” Ruth Porat, president and chief investment officer at Alphabet Inc., told Bloomberg News at the event. But, she added, “it isn’t a foregone conclusion” the US holds on to its advantage. Even that may have been overly optimistic.  That same week, a little-known Chinese AI startup called DeepSeek released a new open-source AI model called R1 Bloomberg 28 January 2025 https://www.bloomberg.com/news/articles/2025-01-28/deepseek-ai-platform-brings-a-1-trillion-market-reckoning-on-cost   Apple will be cheering that its decision not to throw billions at building AI capabilities has paid off.  It can sit back and pick the best models from a newly commoditised selection.   Moreover, efficiency gains are likely to result in greater use of ai.  The Jevons paradox—the...

Dåliga nyheter för Nvidia. Men för resten av börsen är AI-kraschen överdriven

Det behöver inte betyda att resten av börsen blir förlorare. Visst, de stora techjättarna har investerat enorma belopp i AI som förr eller senare ska ge avkastning.  Men vad som egentligen har hänt är att Deepseek har gjort AI billigare och mer tillgängligt. Det är inte dåligt, vare sig för samhället eller för börsen. Tvärtom är det mycket möjligt att detta kommer att bereda väg för en våg av nya AI-vinnare. Johan Wendel DI 27 januari 2025 https://www.di.se/analys/ai-raset-overdrivet-forutom-for-nvidia/ Maria Landeborn råder hushållen att behålla sina Sverige-fonder. november 25, 2024 https://englundmacro.blogspot.com/2025/01/frida-bratt-konstaterar-att-det-ar.html Tillbaka till Rolfs länktips 28 januari 2025 https://englundmacro.blogspot.com/2025/01/rolfs-lanktips-28-januari-2025.html  

Sam Altman, vd för OpenAI som äger Chat GPT det är ”uppfriskande” att ha en konkurrent.

Bild
  ”Deepseeks R1 är en imponerande modell, särskilt sett till vad de kan leverera för priset”, - det är ”uppfriskande” att ha en konkurrent. Dagens industri/TT 28 januari 2025 https://www.di.se/live/trump-deepseek-ar-en-varningsklocka/ DeepSeek - It’s open source and free for personal use https://englundmacro.blogspot.com/2025/01/deepseek-on-monday-panic-had-spread-to.html Tillbaka till Rolfs länktips 28 januari 2025 https://englundmacro.blogspot.com/2025/01/rolfs-lanktips-28-januari-2025.html

DeepSeek on Monday the panic had spread to Wall Street - It’s open source and free for personal use

Bild
  Google Translate: DeepSeek på måndagen hade paniken spridit sig till Wall Street - Det är öppen källkod och gratis för personligt bruk Nasdaq 100-terminer sjönk med mer än 3% och satte teknikaktier på rätt spår för en 1 trillion dollar rutt. Detta kortsiktiga tillvägagångssätt för att göra affärer, att enbart enorma summor pengar kan hålla konkurrensen i schack, är anledningen till att jättar som Meta Platforms Inc. är i panikläge om DeepSeek, ett kinesiskt företag som har byggt en formidabel AI-modell för ungefär lönen för en enskild AI-chef i USA. Dess genombrott innebär nu en förändring i maktbalansen och en räkning för teknikjättar, som plötsligt inte längre är de garanterade vinnarna av AI. Anledningen är uppenbar.  OpenAI tog nyligen in den största riskkapitalrundan i historien (6,6 miljarder USD), medan Mark Zuckerberg har sagt att Meta kommer att spendera upp till 65 miljarder dollar på AI-projekt i år. Teknikjättar har spenderat hårt på AI, och nu ser det ut som ett...

The börstoppen är nådd? Ekonomibyrån Fredagen den 24 januari

 SvT Ekonomibyrån Fredagen den 24 januari 2024 https://www.svtplay.se/video/jXvyoPn/ekonomibyran/the-borstoppen-is-nadd Tillbaka till Rolfs länktips 27 januari 2025 https://englundmacro.blogspot.com/2025/01/rolfs-lanktips-27-januari-2025.html

DeepSeek 2

Bild
The sudden emergence of DeepSeek calls into question the underpinnings of the rally that’s added $15 trillion to the value of Nasdaq 100 Index companies since the end of 2022. It’s also another lesson in the risks of paying high valuations for emerging technologies in the hopes of reaping life-changing investment gains. The catalyst: the release last week of the latest version of DeepSeek, an AI assistant from a Chinese startup that has rocketed to the top of the download charts in app stores. Pundits are raving about the model, saying it compares favorably to ChatGPT while running on less-advanced chips.  Click here for a QuickTake on DeepSeek. https://www.bloomberg.com/news/articles/2025-01-27/what-is-deepseek-r1-and-how-does-china-s-ai-model-compare-to-openai-meta And the damage goes beyond Big Tech. As often happens in a market frenzy, investors loaded up on risk across the spectrum on the way up, and now they’re dumping broadly on the way down. So Bitcoin is getting whacked to...

Wall Street tech stocks head for $1tn sell-off; DeepSeek

Financial Times 27 January 2025 https://www.ft.com/content/674758d7-ffdf-4b88-bb73-f539b56ac4b1 Deep Seek https://englundmacro.blogspot.com/2025/01/deep-seek.html Tillbaka till Rolfs länktips 27 januari 2025 https://englundmacro.blogspot.com/2025/01/rolfs-lanktips-27-januari-2025.html

Deep Seek

Bild
A Chinese artificial-intelligence company has Silicon Valley marveling at how its programmers nearly matched American rivals despite using inferior chips. Born in 1985, Liang grew up in China’s southeastern province of Guangdong. He went to China’s prestigious Zhejiang University and specialized in machine vision. A few years after graduation, Liang founded High-Flyer with two college friends in 2015.  DeepSeek’s development was led by a Chinese hedge-fund manager, Liang Wenfeng, who has become the face of the country’s AI push. On Jan. 20, Liang met China’s premier and discussed how homegrown companies could narrow the gap with the U.S. DeepSeek Chief’s Journey From Math Geek to Global Disruptor Wall Street Journal 28 January 2025 https://www.wsj.com/tech/ai/liang-wenfeng-deepseek-ai-8bba3bb3 Wall Street Journal 27 January 2025 https://www.wsj.com/tech/ai/china-ai-deepseek-chatbot-6ac4ad33 Under helgen ökade spekulationerna om Deep Seeks senaste produkt, som bedöms vara kostnadsef...

Samtidigt som USA-kohorten kräver lägre ränta

Bild
  Walmart och Costcos värderingar har dragit iväg och lågprisjättarna värderas nu högre än de flesta så kallade Magnificent 7-bolagen.   DI 23 december 2024 https://www.di.se/analys/dyra-detaljhandelsjattar-lockar-inte-till-kop/

Expensive stocks can always get pricier

 Is This Wildly Overvalued Stock Market Doomed? Yes, but Maybe Not Yet Stocks are super expensive on just about every measure. That historically has meant low returns in the long run. Hence the dire 10-year forecasts. But history also suggests no link at all between nosebleed valuations like we have today and returns over the next year. Expensive stocks can always get more expensive, and often do.  AI, rising earnings and American exceptionalism. “Investors are assuming all this capital spending will be worthwhile,” James Mackintosh Wall Street Journal 6 December 2024 https://www.wsj.com/finance/stocks/stock-market-overvalued-forecasts-2025-e073e1d4 Tillbaka till Rolfs länktips 6 december 2024 https://englundmacro.blogspot.com/2024/12/rolfs-lanktips-6-december-2024.html