All news

PPhoenix Tech17:49

AI score65

Anthropic Warns of Recursive Self-Improvement Risk in S-1/A Filing

In the risk disclosures of its S-1/A prospectus supplement updated on Monday, Anthropic stated that the company may soon reach the state of recursive self-improvement as defined in its Responsible Scaling Policy. The filing said the company has already used AI to develop new models and achieve some capability jumps, but the team did not notice at the time, and may in the future only notice after an acceleration in R&D has already occurred. Anthropic also stated that none of its existing models have crossed this threshold, and committed to retaining human involvement in frontier model R&D for the foreseeable future, and is considering revising its Responsible Scaling Policy.

Same story · 6 reports

#Safety / alignment#Anthropic#Industry news

PPhoenix Technology17:49

AI score62

Oracle Raises $20 Billion for AI Expansion, Shares Fall Nearly 9%

Oracle announced that it would raise approximately $20 billion through bond issuance and stock sales to support AI and cloud computing infrastructure construction. After the news was announced, the stock price fell 8.8% to $119.22, and the cumulative decline for the year widened to about 39%. The offering consisted of 86.2 million shares priced at $116 per share, a discount of about 11.2% to last Friday's closing price; the bond portion consists of five tranches of fixed-rate notes with maturities ranging from 2029 to 2046 and interest rates between 7% and 8.65%. As of August 31, Oracle's notes payable and other borrowings reached $125.34 billion, and the company expects revenue to exceed $225 billion by fiscal year 2030.

Same story · 2 reports

#Industry news#English text is being prepared.#OpenAI

4404 Media17:42

AI score53

AI scrapes AO3 fanworks to build a site for sale, authors send 2,000 takedown requests in one day

A website called ao3.life, which used AI agents to scrape AO3 fanworks and convert them into paid audiobooks, shut down one day after launching; the creator said it received more than 2,000 takedown requests from authors that day. The site claimed it used a group of AI agents controlling 40 virtual machines to scrape works from the internet, offered the first 3 minutes free, then charged for AI-generated audiobooks, and claimed authors could receive 91% of the revenue. The creator described himself as a 17-year-old developer from India and said the site had 126,592 unique users within 24 hours, but actually sold only about 11 audiobooks.

#English text is being prepared.#English text is being prepared.#Agents

Tt3n17:37

AI score67

OpenAI claims AI solved nearly 400 math problems; mathematicians criticize it for prioritizing results over process

OpenAI claims that with the help of an as-yet-unreleased AI model, it published more than 700 manuscripts in October, covering more than 370 math problems, with each taking an average of three hours, including work related to the Riemann Hypothesis. Institutions such as the German Mathematicians Association criticized this approach as a results-oriented "power display," arguing that the value of mathematical research lies in the proof path and new methods rather than the answers themselves, and expressing concern about the use of unverified knowledge and issues of academic norms. Stuttgart professor Christian Hesse, on the one hand, called OpenAI's demonstration "stunning," and on the other hand, believes that doctoral training and the definition of mathematical understanding need to be readjusted.

#English text is being prepared.#OpenAI#Papers/Research

Tt3n17:37

AI score60

Anthropic's internal AI usage spending exceeds labor costs for the first time

Anthropic has published data for the third quarter of 2026, stating that since January 2026 its internal AI usage has grown 250-fold, and that at list prices, internal AI usage spending has exceeded all labor costs—though this comparison excludes stock-based compensation. The cost is calculated at external list prices rather than the actual cost of self-used infrastructure, and Anthropic has not separately disclosed the total spending on internal AI usage. Anthropic also says that in September 2026, 90% of contributions to its codebase were completed by AI, per-developer code output grew 45-fold compared with 2024, and headcount over the same period grew more than fivefold compared with the end of 2024.

Same story · 5 reports

#Industry news#Anthropic

SSemiconductor Industry Review16:52

AI score67

Report: DeepSeek's Next-Generation Large Model V5 Plans to Use Ascend Chips for Pretraining

Reports suggest that DeepSeek's next-generation large model V5 will be pretrained on Huawei Ascend chips; previously, Ascend was mainly used for inference tasks on DeepSeek models. On September 30, DeepSeek open-sourced the TileLang compilation tool for the Ascend platform, computing libraries such as DeepGEMM, and the DeepEP communication library, corresponding one-to-one with NVIDIA platform components. The article also mentions that teams including Shenzhen Hetao College completed full-parameter post-training of the 1.6-trillion-parameter DeepSeek-V4-Pro on an Ascend 910C cluster, but this did not involve foundational pretraining.

#Industry news#English text is being prepared.#English text is being prepared.

SSemiconductor Industry Watch16:52

AI score55

onsemi to Acquire Ferric for $600 Million as AI Chip Power Delivery Moves Inside the Package

onsemi announced it will acquire integrated voltage regulator (IVR) vendor Ferric for $600 million in cash, plus up to $175 million in contingent consideration tied to future revenue targets, for a total transaction value of up to $775 million. On the same day, onsemi raised its 2030 AI data center revenue target by $500 million to more than $3 billion. Ferric's IVR integrates the power controller, power switches, inductors and capacitors into a single package, reducing board footprint by 50% to 70%; its latest product, the Fe1767 Newport, delivers 145 amps within 49.8 square millimeters at up to 90.5% efficiency, reducing power delivery network impedance and transient voltage droop by up to 57%.

#Industry news#English text is being prepared.

Hheise online16:50

AI score67

Marvell plans to acquire Lumentum for approximately $115 billion

Marvell Technology announced it will acquire optoelectronic components manufacturer Lumentum for approximately $115 billion (€103 billion), with the transaction to be conducted entirely through a stock swap, whereby Lumentum shareholders will receive 4.5 Marvell shares per share; the deal is expected to close in the first half of 2028 and still requires shareholder and regulatory approval. The two companies said the merger will enable them to address the increasingly complex connectivity needs of next-generation AI infrastructure, and that Marvell's serviceable market will expand to $550 billion. In March of this year, Nvidia had already invested $2 billion in each of the two companies; according to Bloomberg, Nvidia CEO Jensen Huang has expressed support for the merger; after the news was announced, Marvell and Lumentum shares fell nearly 13% and nearly 19%, respectively, on Monday.

Same story · 9 reports

#Industry news#English text is being prepared.

Hheise online16:50

AI score57

404 Media reports Proton internal Slack leak: CEO accused of prioritizing growth over security

After interviewing eight current and one former Proton employee and reviewing internal Slack messages and meeting recordings, 404 Media reported that CEO Andy Yen is accused of internally pushing for heavy use of AI to write code in order to accelerate growth, while taking a hard line with employees who raised security concerns. A meeting recording shows that about 90% of Proton's code was written by AI last October; Yen said in a security channel that requiring testing before launch is loser thinking. Proton's CTO subsequently published an AI manifesto, saying that failing to follow the AI trend would threaten the company's mission, and the full text mentions security only when referring to not discussing it in depth.

#Industry news#Safety / alignment#English text is being prepared.

UUnite.AI16:50

AI score63

OpenAI Plans New Security Sprint and Raises Bug Bounty

OpenAI announced plans to conduct a new round of internal security sprints with external researchers such as Cathedral. On the same day, the bug bounty update raised the median reward ranges for medium-, high-, and critical-severity vulnerabilities by an average of more than fourfold. Critical vulnerabilities in core products such as ChatGPT, Codex, or the API Platform can earn up to $30,000, and particularly severe findings can reach $172,000 after bonuses. In previous sprints, OpenAI committed 25% of its product engineers, used Codex to reproduce vulnerabilities and prepare patches, mobilized more than 250 people covering over 100 service areas, and closed 53 urgent or high-priority issues on the first day. OpenAI will also open its strongest models to Daybreak customers for authorized defensive work, a program corresponding to a $1 billion commitment and a six-month usage target.

Same story · 2 reports

#Safety / alignment#OpenAI#Industry news

EEast Money16:50

AI score73

Anthropic's Q3 operating profit turns to a loss as nearly $900 billion in compute spending weighs on results

Anthropic submitted a revised S-1 prospectus to the U.S. SEC, adding Q3 data: revenue of $16.956 billion, up about 12-fold year-over-year, with GAAP operating results swinging from a profit of $974 million in Q2 to a loss of $567 million. Q3 spending on chips, power and data center infrastructure reached $9.9 billion, up 77% quarter-over-quarter, with compute costs rising to 58% of revenue; as of the end of September, total committed spending reached $895 billion. The company said roughly $5 billion in new short-term compute contracts and about $2 billion in non-cash charitable matching expenses weighed on profit; excluding these, adjusted operating profit was $2.5 billion, with margin falling from 19% to 15%.

Same story · 21 reports

#Industry news#Anthropic#OpenAI

TThe Guardian Technology16:48

AI score59

UK Crown Prosecution Service prosecutor criticized by High Court for citing fictitious case law from AI search

A senior UK Crown Prosecution Service prosecutor cited a nonexistent case, CPS – R v J [2024] EWCA Crim 3529, in case documents involving a minor, and two High Court judges called the explanation "unsatisfactory." According to The Guardian, the citation came from an AI summary generated in an internet search; the CPS said it had corrected the error and launched an internal review, stressing that its AI policy prohibits using AI for legal decisions and requires human verification. The matter brought the UK AI-related error tracking entry count to 101, and maintainers said the identified cases may be only the tip of the iceberg.

#Safety / alignment#Industry news

MMediaNama16:48

AI score60

TRAI allows telecom customer service to use AI agents and requires 24×7 service

The Telecom Regulatory Authority of India (TRAI), through the November 2, 2026 amendment to the complaint handling regulations, allows telecom operators to provide customer service using AI agents, automated systems, or a hybrid of human and automated methods, but complaint centers must shift to 24×7 operation, with human assistance retained for at least 16 hours per day. The new rules also require an online feedback survey after each complaint is resolved and quarterly publication of the results, quarterly reports submitted to the CEO or board of directors, the establishment of a grievance oversight committee including representatives of consumer organizations, and a reduction of the grievance window from 30 days to 15 days along with the elimination of the advisory committee step. The report notes that the regulations do not define the specific meaning of "appropriate human oversight," nor is accountability clear when AI agents make errors.

#English text is being prepared.#Agents#Speech

CCaixin16:47

AI score58

Anthropic's Q3 revenue grew 1206.3% year-over-year, with operating profit turning to a loss of $567 million

Anthropic's updated prospectus shows that revenue for the first three quarters of 2026 grew 1206.3% year-over-year to $33.242 billion, while operating loss narrowed 75.9% year-over-year to $1.260 billion. In the second quarter, it had recorded an operating profit of $974 million, but in the third quarter an operating loss of $567 million reappeared, and net loss expanded 314.4% year-over-year to $104.039 billion. Caixin reported that the decline in year-over-year revenue growth and continued increases in computing power spending are the backdrop for the swing to a loss.

Same story · 21 reports

#Industry news#Anthropic

CcnBeta16:47

AI score55

OpenAI Hires Former U.S. Army Secretary Driscoll to Lead National Security Modernization Affairs

OpenAI has hired Driscoll, the 26th U.S. Secretary of the Army, to oversee matters related to national security modernization. He will work with company management and external partners to advise on how advanced artificial intelligence can strengthen security capabilities, and to promote the adoption of related technologies by U.S. allies, while also helping internal teams understand the actual needs of national security agencies. He will also participate in OpenAI's so-called "democratic AI" work; previously, the company had advanced cyber defense, critical infrastructure protection, and public service security through the Daybreak project.

Same story · 7 reports

#Industry news#OpenAI

LLeitech15:56

AI score56

Xiaomi Wins Two Awards at World Internet Conference, Discloses Progress on Embodied Foundation Model in Factory and Home Scenarios

Xiaomi Group received the "Outstanding Contribution Award" at the 2026 World Internet Conference, while its "Key Technologies for a General-Purpose Robot Manipulation Foundation Model Oriented Toward Large-Scale Deployment" won the "Leading Science and Technology Award." Xiaomi also disclosed that its humanoid robot, at the center console side cover panel sequencing workstation in an automobile factory, took on an 8-hour work task for the first time, achieving autonomous correction of anomalies and system-level coordination with the production line; an Agent system built on the Xiaomi-Robotics-1 pre-trained model is in trial operation at Xiaomi Youth Apartments, completing validation of long-horizon tasks lasting more than 30 consecutive minutes, covering 3 rooms, and including 50+ tasks. The model was pre-trained on 100,000 hours of real-world manipulation data, and Xiaomi says it ranks first globally on evaluation leaderboards such as RoboCasa365 and RoboDojo.

Same story · 2 reports

#Industry news#Embodied AI#Agents

EEast Money15:49

AI score80

DeepSeek V5 plans to pretrain on Ascend chips

According to sources in the industry chain, DeepSeek's next-generation large model, DeepSeek V5, plans to conduct pretraining based on Huawei Ascend chips; previously, Ascend was mainly used for inference tasks of DeepSeek models. The report mentions that after the release of DeepSeek-V4 in April 2026, Huawei stated that the full range of Ascend supernode products could fully support the V4 series models; on September 30, DeepSeek open-sourced infrastructure components for the Ascend computing platform, covering TileLang compilation tools, DeepGEMM, FlashMLA, DeepEP, and others. Currently, the implementation of domestic computing power remains concentrated on the inference side; the full-parameter post-training of DeepSeek-V4-Pro previously completed on the Ascend 910C cluster did not involve foundational pretraining.

Same story · 3 reports

#Industry news#English text is being prepared.#DeepSeek

PPhoenix Technology15:48

AI score71

Report: DeepSeek V5 Plans to Use Huawei Ascend Chips for Pretraining

According to a report by the STAR Market Daily, multiple industry chain sources say that DeepSeek's next-generation large model, DeepSeek V5, plans to conduct pretraining based on Huawei Ascend chips. Previously, Ascend chips were mainly used for inference tasks of DeepSeek large models.

Same story · 3 reports

#Industry news#English text is being prepared.#DeepSeek

CcnBeta15:46

AI score55

UK NCSC Warns AI Will Amplify Cyber Threats to Legacy Systems

The NCSC, part of the UK's Government Communications Headquarters (GCHQ), has released a report concluding that it is highly likely hostile actors will adopt artificial intelligence technology, enabling them to carry out large-scale ransomware attacks and targeted social engineering fraud at lower cost in the future and to quickly discover and exploit vulnerabilities. NCSC Chief Executive Richard Horne said new technologies are amplifying fundamental security weaknesses that organizations have long neglected. The report notes that some models from Google, Anthropic and OpenAI have already demonstrated the ability to find software vulnerabilities and analyze code flaws. The UK government also announced investment plans, including 60 million pounds for the first phase of the National Cyber Shield project, 70 million pounds to establish the National Exposure and Real-time Vulnerability Centre (NERV), and the creation of an AI cybersecurity centre of excellence in Manchester operated by the NCSC.

#Safety / alignment#Industry news#OpenAI

CCailian Press14:57

AI score75

DeepSeek V5 plans to conduct pre-training based on Huawei Ascend chips

According to multiple industry chain sources, DeepSeek's next-generation large model DeepSeek V5 plans to conduct pre-training based on Huawei Ascend chips; previously, Ascend was mainly used for inference tasks of DeepSeek models. The report outlined that around the Spring Festival of 2025, Ascend completed the large-scale deployment of DeepSeek-V3/R1 in Huawei Cloud's Gui'an Data Center; after the release of DeepSeek-V4 in April 2026, the full series of Ascend SuperNode products could already support the V4 series; and on September 30, DeepSeek also open-sourced infrastructure components such as TileLang, DeepGEMM, and DeepEP for the Ascend platform. The report also noted that the implementation of domestic computing power is currently still concentrated on the inference side, and the ecosystem migration cost on the training side is the variable that determines the market ceiling.

Same story · 3 reports

#Industry news#English text is being prepared.#DeepSeek