news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

before one wave subsides, another wave rises!

just as nvidia triggered a sharp drop in u.s. technology stocks, another dispute has begun. california's efforts to regulate artificial intelligence have caused huge divisions in the technology community. california's new bill will require artificial intelligence developers to follow certain rules before developing models. the bill has been passed by the state senate, and the state legislature must approve the bill by august 31 before submitting it to governor gavin newsom.

some have tried to suppress this single state's overreach, while others have supported the bill. so far, u.s. legislation on artificial intelligence remains at the state level, and the u.s. congress has not taken action. but since european regulators have already taken a leading role in ai legislation, action by the u.s. congress is also to be expected.

it is worth mentioning that after nvidia's quarterly earnings report, its stock price fell nearly 7% after the market, and its market value evaporated by $200 billion. nasdaq futures fell by about 1%. us technology stocks were generally under pressure. broadcom shares and advanced micro devices both fell by about 2%. microsoft and amazon both fell by nearly 1%. amd, the "big monster stock" of artificial intelligence, plunged 26% during the session. after hindenburg released a short report, the company postponed the release of its earnings report.

in addition, affected by nvidia, cpo concept stocks fell at the opening, with xinyi sheng falling more than 9%, zhongji xuchuan falling more than 8%, tianfu communication falling more than 7%, and shengyi electronics, mingpu optoelectronics, taichen optoelectronics, accelink technology, and huagong technology following suit.

california's actions

artificial intelligence has suddenly emerged since the second half of 2022. at the same time, the supervision of artificial intelligence has also been gradually strengthened.

according to reuters, the california state assembly passed the frontier ai model safety innovation act (sb 1047). the bill is one of the first important ai regulations in the united states. the bill requires ai companies operating in california to take a series of precautions before training complex basic models. these measures include being able to quickly and completely shut down the model, ensuring that the model is protected from "unsafe post-training modifications", and maintaining a testing program to assess whether the model or its derivatives are particularly prone to "causing or contributing to serious harm."

there are no longer any criminal penalties in the bill, but civil penalties remain; the attorney general cannot seek civil penalties until damage has occurred; the frontier models unit, which was originally a new regulator specifically for artificial intelligence, no longer exists; the legal standard for ensuring developers comply with the bill has changed from providing "reasonable assurance" of safety to a lower standard of "reasonable care"; if open source developers spend less than $10 million to fine-tune an existing model, they are not considered "developers" under the bill and the liability will be borne by the original developer; this move is intended to protect open source developers from burdens.

senator scott wiener, the bill's lead author, said sb 1047 is a very reasonable bill that requires large ai labs to do what they have already promised to do: test their large models for catastrophic safety risks. "we have been working hard all year to refine and improve this bill with open source advocates, anthropic (an ai company), and other organizations. sb 1047 is well aligned with what we know about foreseeable ai risks and is worth enacting."

major differences

anthropic expressed cautious support for the bill, while openai opposed it, saying it could "stifle innovation."

former house speaker nancy pelosi and several other california lawmakers urged their home state to reject the bill, and several industry and business groups also expressed opposition, including the u.s. chamber of commerce and the software and information industry association. in a letter first shared with axios, the tech advocacy coalition progressive chamber and other organizations (including the bay area council, context fund, engine, net choice, rstreetinstitute, silicon valley leadership group and tech freedom) wrote to newsom urging him to veto the bill, saying that even with recent changes, the bill is still "fundamentally flawed and misplaced." san francisco mayor london breed announced on tuesday that she also opposed the bill's passage.

critics such as openai also say the federal government is best suited to regulate ai, rather than states. supporters of the proposal point out that congress has not acted so far. and if congress does act in the future, it might pass legislation that preempts state laws.

however, tesla ceo musk expressed support for the bill. he said, "california should pass sb 1047 artificial intelligence safety act," and he believes that the risks of artificial intelligence to the public deserve regulation. "this is a difficult decision that will make some people uneasy." but analysts believe that musk's motivations may be more complicated than a simple desire for regulation. he has previously called for a six-month moratorium on artificial intelligence development, while his own artificial intelligence work continues, and often with fewer safeguards than his competitors.

nvidia's signal

it is worth mentioning that the market's attitude towards nvidia, the leader in artificial intelligence, is also changing. the company expects an adjusted gross profit margin of 75% in the third quarter, with a fluctuation of 50 basis points. according to data from the london stock exchange, analysts on average predict a gross profit margin of 75.5%. nvidia's stock price plummeted nearly 7% after the market.

nvidia's impact is certainly not limited to itself. broadcom, amd, microsoft and other stock prices also followed suit. this morning, tsmc also opened with a drop of more than 2%. wall street has begun to worry that the already huge spending of companies such as microsoft and alphabet will increase further in order to compete with other major players for the dominance of emerging artificial intelligence technology. since the release of the report last month, the share prices of microsoft and alphabet have continued to fall.

last night, the stock price of amd, which has close cooperation with nvidia, plunged more than 26% during the intraday trading session after the company said it would delay the disclosure of its full-year financial report. on tuesday, hindenburg, a well-known short-selling agency, released a short-selling report on the company, saying that after investigation, it found that there were obvious red flags in amd's accounting operations. amd's main business is divided into three categories: servers and storage systems, server software management solutions, and global support services. since mid-2022, the stock has seen a nearly 30-fold increase thanks to the ai ​​boom.

well-known analyst ming-chi kuo said that nvidia's stock price performance after the release of its financial report is more like reflecting short-term investors' expectations for the second quarter financial report, and does not represent the stock price trend in the next few months. the blackwell chip will contribute to revenue in the fourth quarter, which can be regarded as clarifying investors' concerns about the delay in gb200 shipments, which is beneficial to the stock price trend in the next few months.

in response to the outside world's concerns about the company's blackwell chip, nvidia ceo huang renxun said that blackwell chip samples have been sent to all parts of the world and mass production has begun. when the production momentum is up, blackwell will be in large supply and its functions are very good. nvidia will have very brilliant results next year.