Is the tech stock frenzy cooling off?| AI weekly review

1. AI concept stocks fell last week most, from now on to pay attention to the Federal Reserve

2.Investors may be fleeing U.S. technology stocks, AI industry "wave of reduction" has also appeared

3.against Microsoft Google, Amazon cloud smashed 100 million dollars to launch a generative AI innovation center

4.AI may enhance than to replace existing industry vertical applications, Adobe may benefit.

5. "China's most beneficial AI stocks"! Baidu's rating is raised by Morgan, saying "it can rise 31% more

6.AI AppStore is here! Report: OpenAI is considering the launch of AI application store

7.The latest ranking of the big model, GPT-4 tops the list, the new 33 billion parameters "Vicuna" ranked first in open source

After a sustained rally, a pullback in AI stocks is to be expected. Of course, some stocks have pulled back a bit more than expected.

On June 22, at a hearing held by the Senate Banking Committee in Congress, Powell stated that the Federal Reserve's decision last week to maintain the benchmark interest rate was an attempt to slow down the pace of easing, he emphasized that if economic indicators meet expectations, even if policy rates have reached appropriate levels, the Fed may take action again this year to hike rates. This statement caused panic in the market over fears of recession, resulting the technology stock plunged.

Chris Harvey, head of equity strategy at Wells Fargo Securities, said recently that the strong rally in technology stocks will not stop until the Fed becomes more aggressive and destroys the economy.

And now, it is time to pay attention to the Fed.

The market currently has two viewpoints on the Fed's future rate hike expectations. The first view: the Fed's forecast to raise interest rates to 5.50%-5.75% is just inflation expectation management, the future will not continue to raise interest rates, the economy will be in a weak recession at high interest rates to reduce inflation to around 2%.

The second view: The Fed will continue to raise interest rates in the second half of the year. This is due to the possibility of a rebound in nominal inflation growth in the third quarter after a significant reduction in the previous period, while the core growth rate is decreasing too slowly and it is necessary to raise interest rates again continuously to control the risk of inflation rebound.

We can only pay close attention to the performance of the economic data and see the situation as we go. For example, the U.S. core PCE data for May, which will be released on Friday, is worth focusing on.

Investors are fleeing American tech stocks amid concerns over the persistent decline in global asset prices caused by worries about the stability of U.S. monetary policy.

According to Michael Hartnett,head of research at Bank of America, said certain market indicators reveal that investors have begun to shun American technology shares. Citing EPFR Global data, Hartnett notes that although $19 billion flowed into tech stocks in the past two months, there was a $2 billion withdrawal from tech stocks last week alone – representing the largest pullback in ten weeks.

Hartnett also expects that the probability of market downside is significantly higher than upside this summer. He believes the S&P 500 have a maximum upside of 100-150 points and a downside of 300 points by early September.

Furthermore, EPFR Global data shows that last week saw net fund flows withdrawn from global equity mutual funds totaling $5 billion, accompanied by an increase of $5.4 billion invested in bond funds, indicating capital flight from risky assets towards low-risk ones.

U.S. stock market has surged too quickly since October last year, rising more than 20 percent in a sustained period, and thus merits a gasp of breath.

The SEC disclosed that Harvey Jones, a director of NVIDIA, reduced his holdings by 119,795 shares worth $48,297,000 on June 13, and Tench Coxe, a director of NVIDIA, reduced his holdings by 50,000 shares worth $21,107,000 on June 14. Mark Stevens, a director of INVISTA, reduced his holdings by 118,602 shares valued at $51,121,000 on June 16.

It's a small investment for a company with $64 billion in cash reserves and $5 trillion in annual operating expenses. But the announcement shows that Amazon AWS has recognized the importance of generative AI.

Adam Selipsky, CEO of AWS, said in an interview, "You ask yourself this question: In a 10K race, where are the different athletes running the three steps? Does it really matter? The point is, you've run three steps and it's a 10K race.

As part of the latest announcement, Amazon noted that the company will increase the salaries of some data scientists, engineers and solution architects. aws said the center has partnered with companies such as Highspot, Twilio, RyanAir and Lonely Planet.

Amazon has beaten Microsoft and Google in the business of leasing servers and data storage to companies and other organizations to take the lead in the cloud infrastructure market. However, while Amazon has used AI extensively over the years to display shopping recommendations and operate its Alexa voice assistant, its rivals have made greater strides in generative AI.

Selipsky said he is not worried. AI will be the next innovation boom in cloud technology, which will inspire users to adopt cloud technology more aggressively because they need the cloud to enable the use of generative AI. In this regard, AWS offers a unique level of trustworthiness that no other competitor can match. I can't count the number of Fortune 500 companies that have banned ChatGPT because, at least in its initial version, it didn't adequately account for enterprise-level security needs," he said.

Jefferies analyst John Colantuoni recently raised his price target on Amazon from $135 to $150. It said

We think Amazon is still in consensus upside, and the positive sentiment on AI will help revalue Amazon.

Today, it appears that while Amazon is lagging behind other large technology companies in terms of generative AI capabilities, the opportunities presented by AI are just beginning and we expect Amazon's experience in innovation to help them close the gap in the AI space over time.

Loop Capital analyst Rob Sanderson continues to be optimistic about Amazon, raising his price target for the company from $140 to $180, the highest price target given by any investment bank. Morgan Stanley maintained its price target for Amazon at $150 per share.

In early June, OpenAI CEO Sam Altman had a closed-door discussion with the CEO of AI company HumanLooop in London. According to the leaked minutes of the conversation compiled by Geek Park, Sam Altman talked about the future commercialization path of OpenAI:

ChatGPT plugins are exciting for many developers, but Altman believes that they won't be released anytime soon. With the exception of the Brosing plug-in, usage of other plug-ins suggests that PMF (Product/Market Fit) is not yet available. He noted that many people think they want their apps to be in ChatGPT, but what they really want is for ChatGPT to be in the app.

That is, previous market assumptions about vertical applications being replaced by AI may not be very valid, and Adobe's turnaround story confirms this.

The powerful combination of AI+creativity and digital experience applications presented by Firefly can help users improve the efficiency of design, break through the boundaries of creativity, and sustain quality creative energy in the midst of burgeoning content demand, thus filling the gap between imagination and canvas.

This is something that pure AIGC applications cannot do, and the latter does not have its vast user base and user-friendly creative assistance capabilities.

With Adobe's PDF reading software Acrobat Pro the first to raise prices by 40%, some market analysis said that with the addition of AI features, more Adobe software will soon raise prices.

BMO Capital will be Adobe's stock rating from "flat with the market" to "outperform", and its target price from $500 up to $570. The bank believes that Adobe can gain new users through generative artificial intelligence, while the recent survey feedback on Adobe Express is positive.

Morgan Stanley raised its rating on Baidu to Hold and raised its price target by $30 to $190, saying Baidu will rise another 31% with the help of AI.

Gary Yu, an analyst at Morgan Stanley, noted:

In addition, Bank of America noted that Baidu has a leading edge in both technology and applications:

According to technology media The Information, the star unicorn, which has launched GPT series of big models, is considering to launch a "big model store" similar to an app store. The store would allow OpenAI customers to sell their custom AI models to other companies.

According to two people familiar with the matter who spoke to The Information, through the Big Model Store, companies could get access to a variety of AI software that has been fine-tuned, such as AI that solves online retail financial fraud and AI that has the latest data and can answer specific market messages.

It's unclear whether OpenAI will charge commissions for these transactions or otherwise generate revenue from the Big Model Store.

Outside of OpenAI, several companies have built their own Big Model Stores. For example, Salesforce and OpenAI's partner Microsoft.

However, OpenAI denied the news, saying it is not developing a Big Model Store. But The Information also reported that OpenAI CEO Sam Altman disclosed potential plans with some developers at a meeting in London last month.

In fact, many companies are expecting a similar service from OpenAI.

Software maker Aquant bought a license from OpenAI and has adapted the GPT Big Model to answer questions about hardware repair using its own proprietary data.

Aquant has about 100 customers, including Hewlett-Packard, Canon, and Johnson & Johnson. Customers pay thousands to hundreds of thousands of dollars per month to access Aquant's chat robots, which are custom-built on OpenAI technology. The service is available on big model stores such as the Salesforce AppExchange and ServiceNow Store.

In addition, educational institutions, including Khan Academy, have developed Khanmigo, an AI tutor based on unique data, and the company's CEO said it is seeking channels to sell the Khanmigo service.

OpenAI has also explored other similar revenue channels before being rumored to be considering launching a "Big Model Store". Earlier this year, OpenAI launched the GPT plug-in, which allows users to query OpenTable, Shopify or Expedia websites through AI, so they can quickly book or purchase products without having to visit those sites.

However, these plug-ins have not exploded in the market as ChatGPT has. According to The Information, in a meeting with developers last month, Sam Altman directly pointed out that GPT plugins are not finding the right market.

The UC Berkeley-led "LLM Qualifying Tournament" has received its first major update!

The number of models participating in the ranking competition has reached 28, and two new evaluation criteria have been added to the ranking.

The results show that the GPT-4 is still in first place and the GPT-3.5 has successfully regained the second place under the new evaluation system.

UC Berkeley's newly released 33 billion parameter Vicuna was ranked fifth, slightly ahead of the 30 billion parameter WizardLM model developed by Microsoft's Chinese team.

In addition, the existing assessment methods on the market have obvious shortcomings in assessing the human preferences of LLM; therefore, this evaluation standard adds the MT-bench test; and the LLM is used to score the LLM.

The MT-Bench is a well-designed benchmark test containing 80 high-quality, multi-round questions. 8 major categories: Writing, Role-playing, Extraction, Reasoning, Mathematics, Programming, Knowledge I (Science Technology Engineering Mathematics), and Knowledge II (Humanities and Social Sciences). Each of these categories has 10 multiple-round questions, for a total of 160 questions.