CALDEIRÃO DE BOLSA

DeepSeek shocks AI world with 'cheap' Chinese chatbot

Chinese artificial intelligence (AI) app DeepSeek has overtaken ChatGPT and other rivals to become the top-rated free application on Apple's App Store in the US, UK and China.

The app has surged in popularity since its launch in January, challenging the widely-held belief that America is the untouchable leader of the AI industry.

It is powered by the open-source DeepSeek-V3 model, which its researchers claim was developed for less than $6m - significantly less than the billions spent by rivals.

But this claim has been disputed by others in the AI space.

After DeepSeek-R1 was launched earlier this month the company boasted of "performance on par with" one of ChatGPT maker OpenAI's latest models - when used for tasks such as maths, coding and natural language reasoning.

Silicon Valley venture capitalist and Donald Trump advisor Marc Andreessen described DeepSeek-R1 as "AI's Sputnik moment", in a reference to the first artificial Earth satellite that was launched by the Soviet Union in 1957.

Advanced chips power the training of AI models like ChatGPT and DeepSeek.

But since 2021 the US government has widened its restrictions on advanced chips being sold to China.

In order to continue their work without steady supplies of imported advanced chips, Chinese AI developers have shared their work with each other and experimented with new approaches to the technology.

This has resulted in AI models that require far less computing power than before. It also means that they cost a lot less than previously thought possible, which has the potential to upend the industry.

Shares in AI-related companies based in the US, such as Nvidia, Microsoft and Meta were down on Monday morning - and the development knocked European share prices.

ASML, the Dutch chip equipment maker, saw its share price tumble by more than 10% while shares in Siemens Energy, which makes hardware related to AI, plunged by 21%.

"This idea of a low cost Chinese version hasn't necessarily been forefront, so it's taken the market a little bit by surprise," said Fiona Cincotta, senior market analyst at City Index.

"So if you suddenly get this low-cost AI model, then that's going to raise concerns over the profits of rivals, particularly given the amount that they've already invested in more expensive AI infrastructure."

And Singapore-based technology equity advisor Vey-Sern Ling told the BBC it could "potentially derail the investment case for the entire AI supply chain".

But Wall Street banking giant Citi cautioned that while DeepSeek could challenge the dominant positions of American companies like OpenAI, issues faced by Chinese firms could hamper their development.

"We estimate that in an inevitably more restrictive environment, US' access to more advanced chips is an advantage," its analysts said in a report.

Last week, a consortium of US tech firms and foreign investors announced The Stargate Project, a company which is putting $500bn into AI infrastructure in Texas.

Who founded DeepSeek?
The company was founded in 2023 by Liang Wenfeng in Hangzhou, a city in southeastern China.

The 40-year-old, an information and electronic engineering graduate, also founded the hedge fund that backed DeepSeek.

He reportedly built up a store of Nvida A100 chips, now banned from export to China. Experts believe this collection - which some estimates put at 50,000 - led him to launch DeepSeek, by pairing these chips with cheaper, lower-end ones that are still available to import.

Mr Liang was recently seen at a meeting between industry experts and the Chinese premier Li Qiang.

In a July 2024 interview with The China Academy, Mr Liang said he was surprised by the reaction to the previous version of his AI model.

"We didn't expect pricing to be such a sensitive issue," he said.

"We were simply following our own pace, calculating costs, and setting prices accordingly."

MarcoAntonio Escreveu:ChatGPT europeu já está disponível e pretende reduzir atraso na IA para EUA e China - Tecnologias - Jornal de Negócios

O modelo constitui 'um passo importante para a Europa poder assegurar a sua soberania na área da Inteligência Artificial' e foi criado para abranger um número alargado de idiomas.

Um advogado, um presidente de uma IPSS e um CEO da área de “venture capital” juntaram-se para criar uma ferramenta gratuita - com base em inteligência artificial - que responde às dúvidas das entidades do setor social.

O modelo constitui 'um passo importante para a Europa poder assegurar a sua soberania na área da Inteligência Artificial' e foi criado para abranger um número alargado de idiomas.

Num painel que se propunha a falar das previsões tecnológicas para 2025, a inteligência artifical monopolizou a conversa. Às grandes potencialidades da tecnologia estão associadas a grandes riscos, mas a palavra de ordem é continuar - mas com cautela.

No palco da Web Summit, Brad Smith comparou o impacto da inteligência artificial ao surgimento da eletricidade no século passado.

O Bytespider está a recolher dados da internert a uma velocidade 25 vezes mais rápida do que o GPTbot, da OpenAI. Nos últimos meses, a ByteDance tem tentando afirmar-se no mercado da Inteligência Artificial chinês.

A elevada digitalização da administração pública e medidas de ambiente regulado levaram a que a Europa registasse o maior aumento anual de ciberataques: 64%. Os dados são do estudo da NNT Data.

AI innovators win Nobel Prize for physics

Geoffrey Hinton, co-recipient of award with John Hopfield, warns of potential dangers of artificial intelligence


John Hopfield and Geoffrey Hinton have won the physics Nobel Prize for formative work on artificial intelligence that has helped drive scientific advances but raised fears about the risks of abuse.

The award highlights the fundamental role the AI field of machine learning now plays in research, because of the amount of data it can process at speed. Hinton, who quit Google last year so he could speak more freely, said he was “flabbergasted” by Tuesday’s honour and spoke of the power and perils of AI.

(...)

Hopfield, a US physicist, devised an artificial neural network to save and recreate patterns.

Hinton, a British-Canadian computer scientist, used Hopfield’s research to build a new network known as the Boltzmann machine. This can be used to classify images or create new examples of pattern types it has learned.

Together they helped “initiate the current explosive development of machine learning”, the Nobel organisers said.

(...)

Machine learning has become part of our daily lives in areas including facial recognition, language translation and medical diagnosis, said Ellen Moons, chair of the Nobel committee for physics.

(...)

Hinton and fellow AI scientists Yoshua Bengio and Yann LeCun won computer science’s top prize, the Turing Award, in 2018. Hinton currently works at Toronto university and Hopfield at Princeton.

A participar na ronda de financiamento contaram-se várias gigantes da tecnologia, como a Nvidia e a Microsoft - que aliás já tinha investido 13 mil milhões de dólares na criadora do ChatGPT.

O objetivo passa por tornar a empresa mais atrativa para os investidores e a alteração implica que Sam Altman poderá vir a receber capital próprio pela primeira vez. Mira Murati, que chegou a estar alguns dias como CEO interina, vai deixar a OpenAI após seis anos.

A central de Three Mile Island protagonizou o pior incidente nuclear em território norte-americano nos anos 1970. Agora, um dos reatores, que esteve em funcionamento até 2019, vai ser reativado para abastecer de energia os centros de dados para Inteligência Artificial da tecnológica.

OpenAI says new artificial intelligence model can ‘reason’

ChatGPT maker OpenAI unveiled a new artificial intelligence (AI) model Thursday that it says can “reason through complex tasks and solve harder problems” than its earlier models.

The latest model, the first in a new series called OpenAI o1, can perform similarly to PhD students on “challenging” benchmark tests in physics, chemistry and biology, the company said in a blog post.

“We trained these models to spend more time thinking through problems before they respond, much like a person would,” OpenAI wrote. “Through training, they learn to refine their thinking process, try different strategies, and recognize their mistakes.”

The company’s chief technology officer, Mira Murati, touted the models as “the start of a new era in AI.”

“We’re no longer limited by pretraining paradigm; now, we can scale through inference compute, opening up new possibilities for capabilities and alignment,” Murati said in a post on the social media platform X.

While the model released Thursday, known as o1-preview, has greater capability for complex reasoning, OpenAI noted that the last model it released, GPT-4o, may still be more effective for common uses.

o1-preview currently does not have certain features that made earlier models useful, such as web browsing and file and image uploads, the company said.

“o1 is still flawed, still limited, and it still seems more impressive on first use than it does after you spend more time with it,” OpenAI CEO Sam Altman wrote on X.

However, Altman added in a separate post that he is “extremely proud of the team; this was a monumental effort across the entire company.”

OpenAI also unveiled o1-mini, “a faster, cheaper reasoning model that is particularly effective at coding” and is 80 percent cheaper than o1-preview.

o1-preview and o1-mini became available to ChatGPT Plus and Team users Thursday, while ChatGPT Enterprise and Edu users will get access to the new models next week.

Você pode apagar ou modificar a memória diretamente aqui na conversa. Se você deseja que eu apague informações específicas ou todas as informações armazenadas, basta me informar o que deseja remover, e eu farei isso para você.

Se estiver falando sobre a memória local ou os dados da sua conta, esses geralmente são gerenciados pelas configurações da plataforma ou serviço que você está usando, e não pela OpenAI diretamente.

Você pode me pedir para listar ou revisar as informações que tenho armazenadas sobre você a qualquer momento. Se você quiser ver o que foi registrado, basta pedir para eu mostrar a memória ou detalhar o que foi armazenado. Por exemplo, você pode dizer "O que você lembra sobre mim?" e eu fornecerei as informações que tenho registradas.

De acordo com a Bloomberg estará a decorrer uma ronda de financiamento que pretende avaliar a empresa que detém o ChatGPT em 100 mil milhões de dólares.

OpenAI said Friday that it disrupted an Iranian influence operation that was using ChatGPT to generate content related to the U.S. presidential election and other topics.

The network known as Storm-2035 used the company’s chatbot, powered by artificial intelligence (AI), to create content, including “commentary on candidates on both sides in the U.S. presidential election,” that was then shared on social media.

OpenAI has banned the accounts from using its services. It emphasized that the operation “does not appear to have achieved meaningful audience engagement,” with the identified social media posts receiving few or no likes, shares or comments.

“Notwithstanding the lack of meaningful audience engagement resulting from this operation, we take seriously any efforts to use our services in foreign influence operations,” OpenAI wrote in a blog post.

“Accordingly, as part of our work to support the wider community in disrupting this activity after removing the accounts from our services, we have shared threat intelligence with government, campaign, and industry stakeholders,” it added.

The operation used ChatGPT to generate long-form articles that were published to five websites posing as progressive or conservative news outlets, as well as to write short social media comments in English and Spanish from accounts on X and Instagram posing as both progressives and conservatives.

The content mainly focused on the conflict in Gaza, Israel’s presence at the Olympic Games and the U.S. presidential election, although some focused on Venezuelan politics, Latinx rights in the U.S. and Scottish independence, according to OpenAI.

The accounts used by the operation interspersed this content with “comments about fashion and beauty, possibly to appear more authentic or in an attempt to build a following,” the AI startup noted.

OpenAI’s disruption of this Iranian influence operation comes after former President Trump’s campaign said last weekend that some of its internal communications were hacked by “foreign sources hostile to the United States.”

Trump’s campaign pointed to a report from Microsoft on Iran’s influence operations targeting the 2024 election, which revealed that Iranian hackers “broke into the account of a ‘high ranking official’” on a presidential campaign in June.

The same report also featured information on Storm-2035, noting that it was “masquerading as news outlets” and “actively engaging US voter groups on opposing ends of the political spectrum with polarizing messaging on issues such as the US presidential candidates, LGBTQ rights, and the Israel-Hamas conflict.”

OpenAI noted in Friday’s blog post that it “benefited from information about the operation published by Microsoft last week.”