OpenAI’s ambitions to challenge industry giants like Google and Bing,  and startups like Perplexity the company has unveiled its latest innovation: SearchGPT. Designed to offer real-time, AI-powered search capabilities, SearchGPT aims to revolutionize the way users interact with search engines, providing timely and relevant answers sourced from the web.

SearchGPT introduces a user-friendly interface reminiscent of OpenAI’s popular chatbot platform, ChatGPT. Users simply type in their queries, and SearchGPT responds with comprehensive information, images, and links to relevant sources. This functionality is enhanced by the ability to ask follow-up questions and explore related searches through a convenient sidebar.

One standout feature is SearchGPT’s location awareness. OpenAI notes that the service can tailor search results based on the user’s general location, such as providing local restaurant recommendations or weather updates.

SearchGPT leverages the power of OpenAI’s latest models, including GPT-3.5, GPT-4, and GPT-4o. Initially launched as a prototype, it is currently accessible to a select group of users and publishers, with a broader rollout anticipated in the future. OpenAI has indicated plans to eventually integrate SearchGPT’s capabilities into ChatGPT, enhancing its functionality with real-time web information.

[In their blog post](https://openai.com/index/searchgpt-prototype/), OpenAI highlighted the promise of their new tool: "Finding answers on the web often demands considerable effort and several attempts to get accurate results. By integrating real-time web information with the conversational strengths of our models, we aim to make the search process quicker and more efficient.

The launch of SearchGPT comes at a time when AI-powered search tools are facing scrutiny over issues like plagiarism, inaccuracies, and the potential to cannibalize traffic to content creators. Notable incidents include Google’s AI suggesting unconventional pizza recipes and Perplexity’s AI allegedly copying content without proper attribution.

OpenAI has taken steps to address these concerns. SearchGPT is designed to prominently cite and link to publishers, providing clear, in-line, named attributions. OpenAI is collaborating with publishers to refine the search experience and offering tools for website owners to manage how their content appears in search results. Importantly, OpenAI clarifies that participation in SearchGPT does not equate to training OpenAI’s generative models with a site’s content.

SearchGPT focuses on search engines, and it's separate from the training of OpenAI's AI models, It can crawl websites on the internet that are not part of the training.

SearchGPT’s release as a prototype allows OpenAI to fine-tune the service based on user feedback and performance. This strategic approach provides a buffer against potential missteps, such as incorrect or misleading search results, by framing the service as an evolving project.

The development of SearchGPT has been closely watched, with reports from sources like The Information and Bloomberg highlighting OpenAI’s efforts to recruit top talent from competitors like Google. This move signals OpenAI’s intention to establish a strong foothold in the search market, leveraging its advanced AI capabilities to offer a compelling alternative to existing search engines.

OpenAI Unveils SearchGPT: an AI-Powered Search Engine

OpenAI and Broadcom are reportedly in discussions about developing a new AI chip. This collaboration addresses the current shortage of graphics processing units (GPUs) essential for running AI models. OpenAI, a leading AI research organization, has been exploring ways to develop AI chips independently, aiming to reduce its reliance on Nvidia and enhance its supply chain capabilities.

OpenAI's CEO, Sam Altman, has been proactive in this initiative, engaging with various stakeholders, including chipmakers, partners like Microsoft, government bodies, and financial backers. His efforts are geared towards boosting the supply of components and infrastructure required to support increasingly powerful AI models. "The limiting factor of AI is capacity: chip capacity, energy capacity, compute capacity. "OpenAI is determined to take an active role in development rather than waiting for others to lead the way, given their frontline position", said a source knowledgeable about OpenAI's strategy.

The discussions with Broadcom, a leading semiconductor manufacturer, highlight the potential role Broadcom could play in developing a new chip tailored for OpenAI. Broadcom has a strong track record in AI equipment sales and has previously assisted Google in developing its Tensor Processing Unit (TPU). The company's expertise in chip design and manufacturing is expected to complement OpenAI's AI research capabilities.

OpenAI has been actively hiring former Google employees involved in developing Google's TPU, indicating its commitment to building in-house semiconductor capabilities. Additionally, Altman has announced plans to raise funds to establish semiconductor manufacturing facilities, potentially partnering with industry giants like Intel, Taiwan Semiconductor Manufacturing Co, and Samsung Electronics.

The development of a new AI chip is anticipated to have significant implications for the AI industry. It promises faster and more efficient processing of AI models, thereby advancing the field of AI research and development. OpenAI's move to become more self-reliant aligns with its broader goal of pursuing artificial general intelligence (AGI) — AI that can outperform humans across a range of cognitive tasks.

Even with substantial backing from Microsoft, which has committed $13 billion, OpenAI would still require additional financial support or commercial partnerships to realize its ambitious plans. "It’s fair to say it requires massive sums of money to do this stuff," mentioned a source knowledgeable about the plans.

OpenAI's initiative also involves ongoing conversations with industry and government stakeholders to increase access to the necessary infrastructure. This includes collaborations with premier chip designers, fabricators, and data centre developers. The company stated that OpenAI is continuously engaging with industry and government stakeholders to enhance access to the infrastructure required, aiming to make the benefits of AI widely accessible.

OpenAI and Broadcom Collaborate to Develop Advanced AI Chip

OpenAI has introduced its latest small AI model, GPT-4o Mini, which promises to revolutionize the AI landscape by offering a more affordable and faster alternative to existing models. Released on Thursday, GPT-4o Mini is now available for developers, as well as for consumers through the ChatGPT web and mobile app. Enterprise users will gain access next week. This new model aims to make AI more accessible and cost-effective, aligning with OpenAI’s mission of democratizing AI technology.

#### Performance and Capabilities

GPT-4o Mini is designed to outperform industry-leading small AI models in reasoning tasks involving text and vision. It replaces GPT-3.5 Turbo as the smallest model offered by OpenAI and is positioned as a superior alternative to models like Google's Gemini 1.5 Flash and Anthropic’s Claude 3 Haiku. The new model achieves impressive scores of 82% on the Measuring Massive Multitask Language Understanding (MMLU) benchmark and 87% on the MGSM math reasoning test, surpassing its competitors in both metrics.

#### Cost Efficiency and Speed

Major advantages of GPT-4o Mini are its cost efficiency and it's 100x cheaper than OpenAI's GPT-3 Davinci model It is significantly more affordable to run than its predecessors, with a runtime cost of more than 60% lower than GPT-3.5 Turbo. For developers using OpenAI’s API, GPT-4o Mini is priced at 15 cents per million input tokens and 60 cents per million output tokens, making it an attractive option for high-volume, simple tasks. The model also boasts a context window of 128,000 tokens, equivalent to the length of a book, and has a knowledge cutoff of October 2023.

In terms of speed, GPT-4o Mini delivers a median output speed of 202 tokens per second, more than twice as fast as GPT-4o and GPT-3.5 Turbo. This rapid processing capability makes it ideal for speed-dependent use cases, including many consumer applications and agentic approaches to using large language models (LLMs).

#### Multimodal Support and Future Enhancements

Currently, GPT-4o Mini supports text and vision in the API, but Sam Altman confirms on X to release two models in a month Alpha and GA including voice capabilities. This multimodal functionality could enable more advanced virtual assistants that understand and respond to complex user needs, such as creating travel itinerary suggestions based on uploaded documents and images.

#### Enterprise Tools and Compliance

In addition to GPT-4o Mini, OpenAI announced new tools for enterprise customers. The Enterprise Compliance API is designed to help businesses in highly regulated industries, such as finance, healthcare, legal services, and government, comply with logging and audit requirements. This API will provide records of time-stamped interactions, including conversations, uploaded files, and workspace users, giving admins greater control and transparency over their ChatGPT Enterprise data.

Moreover, OpenAI is enhancing its granular control features for workspace GPTs, as well as custom versions of ChatGPT tailored for specific business use cases. Previously, admins could only fully allow or block GPT actions within their workspace. Now, they can create an approved list of domains that GPTs can interact with, providing more nuanced control over the AI's functionality.

OpenAI Unveils GPT-4o Mini: A Game-Changer for Developers and Enterprises

Elon Musk announced that he will be relocating the headquarters of his companies, SpaceX and X, from California to Texas. This decision, shared by [Musk on X](https://x.com/elonmusk/status/1813290895334383820), is driven by various factors including new legislation in California that Musk perceives as detrimental to both families and businesses.

Musk's announcement comes on the heels of California Governor Gavin Newsom signing a bill that prohibits public schools from enforcing policies that require parental notification if a child changes their name, pronouns, or gender identity. Musk, who has been vocal about his disapproval of such measures, cited this new law as "the last straw," expressing concerns that families might need to leave California to protect their children. This sentiment echoes his past statements, where he has criticized California's policies and linked them to broader socio-political issues.

Additionally, Musk pointed to the increasing difficulties he faces in San Francisco, where X's headquarters are currently located. He described the challenges of navigating through the city, citing "gangs of violent drug addicts" as a significant concern. This dissatisfaction has likely been brewing for some time, as evidenced by recent reports from the San Francisco Chronicle indicating that X was seeking to sublease its current headquarters.

This isn't the first time Musk has decided to move one of his companies to Texas. In the past, he relocated Tesla's headquarters to the Lone Star State and has previously considered moving SpaceX's business incorporation from Delaware to Texas. The billionaire CEO's actions are part of a broader trend of tech companies reconsidering their base of operations due to various state policies and living conditions.

Musk's decision also reflects his personal experiences and political views. He has had a public falling out with one of his children over gender identity issues, which has likely influenced his stance on related legislative matters. Moreover, Musk has made controversial statements about gender-affirming care for transgender youth, adding another layer to his reasons for the relocation.

The practical implications of these moves are still unfolding. While SpaceX and X's headquarters will shift to Texas, it remains unclear how this will affect the workforce and facilities currently based in California. Notably, despite Tesla's official move to Texas, the company maintains a significant presence in California, underscored by Governor Newsom's appearance at a recent Tesla event in the state.

A push for in-person work has also marked Musk's operational strategies. After acquiring X, he mandated a return to office for most employees, emphasizing a "hardcore" work culture. This move and the recent decision to sublease the San Francisco office space indicate a significant shift in how Musk's companies will function.

Elon Musk Relocates SpaceX and X Headquarters to Texas Over California Policies

AT&T, a leading U.S. telecommunications giant, has disclosed a data breach affecting millions of its customers. The breach, which occurred between May 1, 2022, and October 31, 2022, involved the unauthorized access and theft of call and text-message records. This article provides a detailed analysis of the breach, its implications, and the steps AT&T and Snowflake, the cloud platform involved, are taking in response.

#### Details of the Breach
AT&T confirmed that the breach compromised the phone records of "nearly all" its customers, including both cellular and landline users. The compromised data includes phone numbers, call and text message logs, and cell site IDs. Notably, the content of the calls and texts was not accessed, but metadata such as who contacted whom, the duration of calls, and the total number of calls and texts were included. This metadata can reveal significant information about user interactions and approximate locations.

The breach also affected customers of other carriers that rely on AT&T's network. In total, approximately 110 million AT&T customers will be notified about the breach. Some of the compromised records are more recent, dating from January 2, 2023, although the exact number of these cases is unspecified.

#### Source of the Breach: Snowflake
The breach was linked to Snowflake, a cloud data platform that AT&T and other companies use to analyze large amounts of customer data. The breach was part of a broader spate of data thefts targeting Snowflake's customers. The data was stolen due to the absence of enforced multi-factor authentication (MFA) on Snowflake accounts, a security lapse that Snowflake attributed to its customers.

Snowflake's Chief Information Officer, Brad Jones, stated that investigations by cybersecurity firms Mandiant and CrowdStrike found no evidence of a vulnerability, misconfiguration, or breach of Snowflake’s platform itself. Mandiant identified the cybercriminal group UNC5537 as the perpetrator, describing them as financially motivated with members in North America and Turkey.

#### Response and Mitigation Efforts
Upon discovering the breach on April 19, AT&T reported the incident to the FBI and the Department of Justice (DOJ). The authorities agreed to delay public notification twice due to potential national security and public safety risks. The FBI verified that it is working with AT&T to enhance investigative efforts and manage the incident response. At least one individual has been apprehended in connection with the breach, though they were not an AT&T employee.

AT&T has taken additional cybersecurity measures to close off the point of unlawful access and is working closely with law enforcement to apprehend the cybercriminals involved. The company has assured customers that their names, Social Security numbers, and credit card information were not compromised. However, AT&T cautioned that phone numbers could be linked to individuals through online tools.

#### Previous Incidents and Broader Implications
This breach marks AT&T's second significant security incident this year. In March, the company had to reset the account passcodes of millions of customers after encrypted passcodes were leaked on a cybercrime forum. In February, a technical malfunction led to a significant cellphone outage, impacting over 1.7 million customers.

The recurring security issues highlight the challenges and responsibilities that come with being one of America's leading wireless carriers. AT&T's network supports approximately 127 million connected devices, including 87 million postpaid wireless subscribers, making it a prime target for cyberattacks.

Massive Breach Exposes Call and Text Records of Nearly All AT&T Cell Customers

Microsoft has relinquished its board observer seat at OpenAI, a move echoed by Apple’s decision not to join the board. This comes at a time of increasing scrutiny from antitrust officials in Europe and the US.

#### [Details from the Financial Times](https://www.ft.com/content/ecfa69df-5d1c-4177-9b14-a3a73072db12)
Microsoft, which has invested $13 billion in OpenAI, announced that its withdrawal from the board would be "effective immediately." Apple, initially expected to take an observer role on OpenAI’s board as part of a deal to integrate ChatGPT into its devices, will also refrain from doing so, according to insider sources.

An OpenAI spokesperson highlighted a shift towards hosting regular meetings with strategic partners, including Microsoft, Apple, and investors like Thrive Capital and Khosla Ventures. This new approach, led by recently appointed CFO Sarah Friar, aims to keep key partners informed and engaged without formal board roles.

#### Antitrust Investigations
The backdrop to these changes includes rising antitrust concerns. The European Commission is exploring a potential investigation into the Microsoft-OpenAI partnership, while the US Federal Trade Commission is examining investments by tech giants in generative AI startups.

Microsoft’s deputy general counsel, Keith Dolliver, noted in a letter to OpenAI that the decision to leave the board follows significant progress made by OpenAI’s newly formed board. Dolliver expressed confidence in the company’s direction, making Microsoft's board role no longer "necessary."

An OpenAI spokesperson expressed gratitude for Microsoft's confidence and emphasized the continuation of their partnership.

These strategic shifts are not isolated events. UK regulators began scrutinizing the Microsoft-OpenAI partnership in December, shortly after the turbulent period that saw CEO Sam Altman’s temporary ousting and return. EU regulators are also investigating this partnership along with other AI-related deals by major tech companies.

Microsoft’s substantial investment in OpenAI has established it as the exclusive cloud partner for OpenAI, powering all its workloads across products, API services, and research. This collaboration has given Microsoft a competitive edge in the AI sector, enhancing its Bing search engine, Copilot, and other AI features across its product lineup.

Microsoft and Apple Step Back from OpenAI Board: Strategic Shifts Amid Antitrust Concerns

AI video company Runway is reportedly seeking $450 million in funding, potentially valuing the firm at $4 billion. Private equity firm General Atlantic is said to be leading the investment talks.

Runway's software enables users to generate new videos using text or image inputs, addressing the growing AI-powered video creation tools market. The company previously secured $141 million from investors including Google and Nvidia, with a valuation of $1.5 billion.

This potential investment highlights AI's increasing role in industry transformation and digital innovation. Runway's AI-driven approach to video production positions it as a significant player in the expanding field of artificial intelligence applications.

AI video startup Runway in talks with General Atlantic to raise funds at $4 billion valuation

A newly discovered technique, aptly named "Skeleton Key," has emerged as a significant threat, demonstrating the ability to bypass safety guardrails in prominent AI models such as Meta's Llama3, Google's Gemini Pro, and OpenAI's GPT-3.5. This method has serious implications, enabling users to extract harmful and forbidden information from these models.

Skeleton Key operates through a multi-step strategy designed to force AI models to ignore their built-in safety mechanisms. These guardrails are crucial for distinguishing between benign and malicious requests. By narrowing the gap between what the model can do and what it is willing to do, Skeleton Key manipulates the AI into providing restricted information. This includes instructions for creating rudimentary fire bombs, bioweapons, and other dangerous content.

[Mark Russinovich](https://www.microsoft.com/en-us/security/blog/2024/06/26/mitigating-skeleton-key-a-new-type-of-generative-ai-jailbreak-technique/), Microsoft's Azure Chief Technology Officer, highlighted the destructive potential of Skeleton Key in a recent blog post. Unlike other jailbreak techniques that solicit information indirectly or with encodings, Skeleton Key directly forces the model to divulge sensitive data through simple natural language prompts.

Microsoft's tests on several AI models exposed a widespread vulnerability to Skeleton Key. Models such as Meta's Llama3, Google Gemini Pro, OpenAI's GPT-3.5 Turbo, OpenAI's GPT-4, Mistral Large, Anthropic Claude 3 Opus, and Cohere Commander R Plus were all susceptible to this technique. Notably, OpenAI's GPT-4 demonstrated some resistance, yet the threat remains significant across the board.

Microsoft has implemented software updates to mitigate its impact on its AI models, including the Copilot AI Assistants. Despite these efforts, the existence of such a potent jailbreak technique underscores the ongoing challenges in developing robust safeguards for generative AI systems.

Russinovich advises companies building AI systems to enhance their models with additional guardrails. He emphasizes the importance of monitoring both inputs and outputs to detect abusive content and prevent the extraction of harmful information.

The Skeleton Key Technique: Jailbreaking AI models

Reports from industry analysts suggest that by 2026, Apple will introduce AirPods equipped with infrared (IR) camera modules. While it may seem unconventional for a device primarily designed to deliver audio, there are compelling reasons behind this development, rooted in enhancing spatial experiences.

Apple has consistently emphasized spatial audio in its audio products, including Apple Music. Spatial audio provides a three-dimensional sound experience, making users feel as if the sound is coming from various directions around them. 

According to [Apple analyst Ming-Chi Kuo](https://medium.com/@mingchikuo/%E9%85%8D%E5%82%99ir%E7%9B%B8%E6%A9%9F%E7%9A%84%E6%96%B0%E6%AC%BEairpods%E5%8F%AF%E6%9C%9B%E6%8F%90%E5%8D%87%E4%BA%BA%E6%A9%9F%E4%BA%A4%E4%BA%92%E9%AB%94%E9%A9%97%E8%88%87%E5%BC%B7%E5%8C%96apple%E7%A9%BA%E9%96%93%E9%9F%B3%E8%A8%8A-%E7%A9%BA%E9%96%93%E9%81%8B%E7%AE%97%E7%94%9F%E6%85%8B-new-ir-camera-equipped-airpods-to-enhance-2daa96913d7f), the IR cameras in the new AirPods will detect environmental image changes, enabling more precise head movement tracking. This would allow for richer spatial audio experiences. For instance, when watching a video with the Vision Pro headset and AirPods, if the user turns their head towards a specific direction, the sound from that direction could be amplified, enhancing the realism of the experience.

In addition to improving spatial audio, the built-in cameras will support new features such as in-air gesture control. This means users could interact with their AirPods without physically touching them, simply by using gestures like nodding or shaking their heads. These gestures could be used for various functions, such as answering or declining calls.

Furthermore, Apple plans to introduce Personalized Spatial Audio with head tracking for gaming. This feature will be available to users with the latest AirPods, AirPods Pro, or AirPods Max, and is set to roll out with iOS 18 this fall.

The new AirPods with IR cameras will likely integrate seamlessly with Apple's broader ecosystem, particularly in the realm of spatial computing and augmented reality (AR). By combining these AirPods with devices like the Vision Pro headset, users will experience an enhanced interaction between audio and visual content in physical space. 

Apple plans to start mass production of these innovative AirPods by 2026, with major supplier Foxconn expected to produce the IR camera modules.

Apple AirPods: The Future of Spatial Audio with Built-In Cameras

Runway has once again pushed the boundaries of what's possible with its latest release: the Gen-3 Alpha text-to-video model. This groundbreaking technology represents a significant leap forward in AI-generated video, offering unprecedented levels of quality, control, and creative potential.

Runway has been at the forefront of text-to-video AI technology since its inception, with its previous models, Gen-1 and Gen-2. Gen-3 Alpha is significantly improved. Gen-3 Alpha delivers a level of accuracy, consistency, and motion.

Gen-3 Alpha excels at producing videos with lifelike movements, seamless transitions, and an impressive level of detail.

While previous generations were limited to 2-3 second clips, Gen-3 Alpha can generate initial video clips of 10-11 seconds, allowing for more complex and engaging content creation.

The new model offers more natural motion and improved realism, addressing earlier versions' limitations.

Users have access to a range of customization tools, allowing them to fine-tune generated videos to their exact specifications. This includes adjusting character features, modifying scene elements, and controlling camera angles and movements.

Currently, Gen-3 Alpha is available to a select group of Creative Partners Program members. To use the model, users will need a paid Runway account.

Runway Gen-3 Alpha is openly available

As generative AI makes its mark across various sectors, the companies manufacturing the necessary chips are reaping substantial rewards. Nvidia, in particular, has a commanding presence in the AI chip market, controlling an estimated 70% to 95% of the market share. Giants like Meta and Microsoft are investing heavily in Nvidia GPUs to maintain their competitive edge in the AI arena. However, this dominance isn't without its challenges. Many generative AI vendors, reliant on these chips, are seeking alternatives. This has led to a surge in interest from venture capitalists and AI vendors in finding new players to rival Nvidia's dominance.

One of the most promising new entrants is Etched, a startup founded just two years ago by Harvard dropouts Gavin Uberti and Chris Zhu, along with Robert Wachen and former Cypress Semiconductor CTO Mark Ross. Based in Cupertino, California, Etched focuses on developing specialized chips for AI models, specifically transformers. Transformers have become the foundation of many generative AI applications, including OpenAI's ChatGPT, Google's Gemini, and Anthropic's Claude.

Etched's flagship chip, Sohu, is an application-specific integrated circuit (ASIC) optimized for transformers. Using TSMC's 4nm process, Sohu is designed to outperform general-purpose GPUs in inferencing tasks while consuming less power. Uberti claims that Sohu is significantly faster and more cost-effective than Nvidia's latest GPUs, with a single Sohu server capable of replacing up to 160 H100 GPUs.

The efficiency of the Sohu chip stems from its streamlined design, which allocates more resources to computational power and less to memory. This is possible due to the predictable data flow in transformer models, allowing Etched to minimize unnecessary hardware components. Uberti likens the potential impact of specialized AI chips to the rise of ASICs in cryptocurrency mining, which dramatically reduces the demand for Nvidia GPUs in that field.

Despite Nvidia's substantial market lead and the challenges inherent in the semiconductor industry, Etched has attracted significant investor interest. The company recently secured $120 million in a Series A funding round, led by Primary Venture Partners and Positive Sum Ventures, with contributions from notable investors like Peter Thiel and GitHub CEO Thomas Dohmke. This brings Etched's total funding to $125.36 million.

Etched is not alone in this endeavour. Other startups, such as Cerebras Systems and Tenstorrent, are also developing innovative AI chips, aiming to capture a share of the expanding market. However, Etched's focus on transformer models sets it apart, and the founders believe their chip will be crucial for future AI applications, including video generation, robotics, and real-time voice interactions.

Etched plans to bring Sohu to market soon, with preliminary commitments already in place from several customers. The company is also launching the Sohu Developer Cloud, an interactive platform that will allow potential clients to test the chip's capabilities.

Although the path ahead is fraught with challenges, Etched's founders are optimistic. They acknowledge the high stakes but believe their specialized approach and rapid development will give them an advantage. As Uberti aptly notes, the future of AI technology depends on scalable infrastructure, and Etched aims to be at the forefront of this transformation.

Etched's Sohu Chip: A New Player in the AI Hardware Market

Boeing’s Starliner spacecraft, on its maiden crewed mission, is experiencing persistent issues that have delayed the return of two NASA astronauts from the International Space Station (ISS). Astronauts Butch Wilmore and Suni Williams, who arrived at the ISS on June 6, were initially scheduled to spend eight to ten days in space. However, their return has been postponed multiple times due to technical problems.

#### Initial Delays and Current Issues
The Starliner’s troubles began even before launch. Initially set for May, the launch was delayed due to a helium leak and power supply issues, eventually lifting off on June 5. Since arriving at the ISS, the spacecraft has encountered five additional helium leaks and problems with its reaction control system thrusters, which are crucial for manoeuvring. These issues have necessitated a thorough assessment and delayed the astronauts' return to Earth.

On June 18, Boeing announced another delay, extending the mission to June 26 to address these issues and accommodate two planned spacewalks. The delay was compounded by concerns about the thrusters and a propellant valve that failed to close completely. As of now, no new return date has been provided.

Steve Stich, NASA’s Commercial Crew Program manager, emphasized that mission managers are taking a cautious approach, ensuring all safety protocols are followed meticulously. The crew, meanwhile, is not in immediate danger and has ample supplies on the ISS. Wilmore and Williams, both veteran astronauts, are well-equipped to handle the extended mission. Wilmore has previously spent 178 days in space, while Williams has logged 322 days.

NASA and Boeing are working closely to resolve the issues. The Starliner’s expendable propulsion system, part of its service module, is at the heart of the current problems. The system is essential for detaching the spacecraft from the ISS and re-entering Earth’s atmosphere. Overheating thrusters and helium leaks related to the frequency of thruster use are being scrutinized through simulations and data analysis in Houston.

The mission management team, comprising NASA and Boeing personnel, is exploring various fixes, including software updates and hardware usage adjustments. If these efforts prove successful, the Starliner will use its thrusters to undock from the ISS and embark on a six-hour journey back to Earth, culminating in a landing in the southwestern United States, aided by parachutes and airbags.

Despite the propulsion system challenges, NASA maintains that the Starliner is still capable of safely returning the astronauts if necessary. In an emergency, the spacecraft could serve as an escape pod from the ISS. However, if the technical issues cannot be resolved, NASA has contingency plans. One option includes using SpaceX’s Crew Dragon, which is currently docked at the ISS and capable of carrying additional astronauts in an emergency.

This mission marks a critical test for Boeing as it seeks to establish itself in the commercial human-spaceflight market, a sector currently dominated by SpaceX. Boeing and SpaceX were selected by NASA in 2014 to develop commercial space transport capabilities. SpaceX’s Crew Dragon has been operational since 2020, whereas Boeing’s Starliner has faced multiple delays and technical challenges.

Elon Musk, CEO of SpaceX, previously criticized Boeing for having "too many non-technical managers," suggesting organizational issues might be contributing to these technical problems. On Earth, Boeing has also faced scrutiny for quality control in its aviation operations, with whistleblowers alleging that the company has cut corners.

Boeing's Starliner Faces Continued Challenges, Delays Astronauts' Return

The AI industry is buzzing with the latest release from Anthropic: the Claude 3.5 Sonnet. Positioned as a significant upgrade from its predecessor, Claude 3 Opus, Claude 3.5 Sonnet boasts remarkable improvements in speed, cost efficiency, and intelligence. This model not only marks a substantial milestone for Anthropic but also sets a new benchmark in the rapidly evolving AI landscape.

#### Speed and Cost Efficiency
Claude 3.5 Sonnet is twice as fast and five times cheaper to operate than Claude 3 Opus, a fact that underscores its technical superiority. In a competitive move following OpenAI's recent release of GPT-4o, Anthropic has made Claude 3.5 Sonnet accessible for free on both the web and iOS platforms, with additional availability to developers through the Anthropic API.

#### Advancements in Intelligence and Functionality
Described by Anthropic product manager Michael Gerstenhaber as “the most intelligent model in the world,” Claude 3.5 Sonnet is designed to excel in understanding nuance, humor, and complex instructions. It delivers high-quality, relatable content and demonstrates proficiency in multiple languages, including German, Italian, Dutch, and Russian, beyond the capabilities of Claude 3.

While Claude 3.5 Sonnet cannot search the internet or generate image files like some competitors, its intelligence is evident through its performance on various benchmarks and qualitative feedback from early users. This feedback highlights the model’s intelligence and humor, likely resulting from Anthropic’s efforts to instill a genuine curiosity and engagement with users' views and values.

#### Artifacts and Collaborative Capabilities
One of the standout features of Claude 3.5 Sonnet is the introduction of “artifacts.” This feature allows users to draft documents, generate code, or design websites in a separate panel that updates in real-time based on user input. This functionality hints at Anthropic’s vision of transforming AI from mere chatbots to collaborative tools, akin to software like Notion or Google Docs.

Moreover, Claude 3.5 Sonnet is being positioned as an essential tool for the workplace. The model supports complex, context-sensitive tasks, making it ideal for customer support and multistep workflows. Anthropic is also planning to enhance this model with web search capabilities and memory functions, which would enable Claude to remember user preferences and interaction history, further increasing its utility in professional environments.

#### Safety and Ethical Considerations
Safety has been a paramount concern for Anthropic. Claude 3.5 Sonnet underwent rigorous safety testing by the UK’s AI Safety Institute before its release, with findings shared with the US AI Safety Institute. The model is classified at ‘AI Safety Level 2,’ indicating it does not pose a risk of catastrophic harm, though it does exhibit increased capabilities in risk-relevant areas compared to previous models.

Anthropic has emphasized that they do not train their generative models on user-submitted data without explicit permission, ensuring a high standard of privacy. This approach is part of their broader strategy to ensure responsible scaling and deployment of AI technologies.

#### Market Position and Future Developments
Anthropic is not resting on its laurels. Following the release of Claude 3.5 Sonnet, they plan to introduce two additional models, Claude 3.5 Haiku and Claude 3.5 Opus, later this year. This continuous innovation aims to enhance the balance between intelligence, speed, and cost every few months, maintaining their competitive edge in the AI gold rush.

The market for AI chatbots is fiercely competitive, with major players like OpenAI, Google, Amazon, and Adobe all vying for dominance. Each company is striving to integrate AI technology seamlessly into both personal and professional environments, building loyal user bases to secure long-term relevance.

Claude 3.5 Sonnet represents a significant leap forward for Anthropic, not just in terms of technical prowess but also in shaping the future of AI as collaborative and safe tools. As the industry continues to evolve, the impact of such advancements will be profound, influencing how we interact with AI in our daily lives and workplaces.

Anthropic's Claude 3.5 Sonnet: A Leap Forward in AI Technology

Ilya Sutskever is a famous AI researcher and Co-founder of OpenAI, a few months ago in OpenAI drama Ilya voted to fire Sam Altman, Sam was returned in a few days and Ilya disappeared for months, [On May 15](https://x.com/ilyasut/status/1790517455628198322) Ilya announced that he is leaving OpenAI company

Today [Ilya announced on X](https://x.com/ilyasut/status/1803472978753303014) that he has started a new company called Safe Superintelligence Ince (SSI), aiming to create a safe, powerful artificial intelligence system within a research organization.

SSI is based in Palo Alto and it has offices in Tel Aviv. Ilya wants to compete with AI startups and continue his AI research passion without distraction. "world’s first straight-shot SSI lab, with one goal and one product: a safe superintelligence." [SSI posted on X](https://x.com/ssi/status/1803472825476587910)

Ilya Sutskever started a new AI company called SSI

Who would have thought that Raspberry Pi, the maker of tiny, affordable single-board computers, would become a public company? Yet, this is exactly what’s happening. On Tuesday morning, Raspberry Pi priced its initial public offering (IPO) on the London Stock Exchange at £2.80 per share, valuing the company at £542 million ($690 million). Following the announcement, shares surged 32% to £3.70, indicating that Raspberry Pi could raise over $200 million during its IPO process.

For now, only certain institutional shareholders can trade Raspberry Pi shares, but retail investors will have access starting on Friday. This listing marks a significant win for the London stock market, which has seen companies like Deliveroo and Wise choose it over the more liquid U.S. stock markets.

Raspberry Pi is renowned for its versatile, ARM-based computers, which have found favor among tech enthusiasts for creating media servers, retro game consoles, interactive dashboards, robotics projects, and more. The company's offerings, the cheapest of which is just $15, provide powerful, programmable solutions that are both cost-effective and energy-efficient.

Beyond the hobbyist market, Raspberry Pi has seen significant uptake in the industrial sector. Many companies now integrate Raspberry Pi units into their devices and facilities, with the industrial and embedded segment accounting for 72% of the company’s sales. Since its inception, Raspberry Pi has sold 60 million units, generating $266 million in revenue and $66 million in gross profit in 2023 alone.

The IPO will provide Raspberry Pi with the capital needed to expand its talent pool and enhance its product offerings. Plans include paying more engineers, bringing aspects of semiconductor design in-house, and introducing new, more expensive products to better serve customer needs. A 2023 fundraising round led by UK chip maker Arm Holdings previously valued Raspberry Pi at around £444 million ($561 million), with estimates now rising to roughly £500 million ($630 million).

Raspberry Pi Ltd is the commercial subsidiary of the Raspberry Pi Foundation, which aims to make coding accessible through low-cost, programmable computers. The Foundation remains the majority shareholder of Raspberry Pi Ltd. Other key shareholders include ARM and Sony Semiconductor Solutions Corporation, with ARM planning to increase its stake via the public listing.

"When we released our first product in 2012, our goal was to provide a computer that was affordable enough for young people to own and explore with confidence," said Eben Upton, Raspberry Pi CEO and founder. With its IPO and expanded resources, Raspberry Pi now has the "technology roadmap to play an increasingly significant role" in the computing industry.

Raspberry Pi: From Hobbyist Favorite to Public Company

Apple is reportedly venturing into the realm of home robotics. Following the recent discontinuation of its much-anticipated electric vehicle project, the company is now exploring personal robots that could revolutionize the way we interact with technology at home. According to [Bloomberg](https://www.bloomberg.com/news/articles/2024-04-03/apple-explores-home-robots-after-abandoning-car-efforts), Apple is in the preliminary stages of developing mobile robots designed to follow users around their homes, potentially marking a significant shift in the company’s product strategy.

Apple’s pivot from its decade-long electric car endeavor, known as Project Titan, to robotics signifies a notable strategic redirection. Despite substantial investments and years of development, the electric car project was officially shelved earlier this year. This shift aligns with Apple’s ongoing quest to diversify its revenue streams and mitigate the slowdown in iPhone sales, particularly as consumers hold onto their devices for longer periods.

Reports suggest that Apple is not merely interested in creating functional robots but aims to integrate advanced artificial intelligence to enhance user interaction and home automation. One of the envisioned products includes a mobile robot capable of autonomous navigation within a home environment. This robot could serve as a personal assistant, enhancing day-to-day activities by following users, assisting with chores, and potentially even participating in video calls.

Additionally, Apple is developing a more advanced tabletop home device featuring a robotic display. This smart display, designed to mimic human head movements, could provide a more interactive and dynamic experience during video calls. The device could also include features like zooming in on specific individuals during a call, thereby enhancing the overall communication experience. However, this project has seen its share of internal debate, with executives divided over its feasibility and market potential.

Apple’s exploration of home robotics comes at a time when the company faces significant challenges. The first half of 2024 has seen declining iPhone sales in critical markets like China, a hefty fine from the European Union, and ongoing antitrust litigation in the United States. These setbacks have fueled criticism that Apple is lagging behind competitors such as OpenAI, Microsoft, and Google in the AI and robotics arena.

By shifting focus to home robotics, Apple aims to regain its innovative edge. The development of AI-driven home devices could position Apple as a leader in a burgeoning market segment. The company’s experience in integrating hardware and software seamlessly provides a solid foundation for creating sophisticated home robots that could surpass existing products in functionality and user experience.

Apple’s interest in robotics dates back to 2019, when the project was spearheaded by Doug Field, who has since moved to Ford as an executive in the electric vehicle division. Initially, Apple’s robotics efforts encompassed a broad range of concepts, including quiet indoor drones. However, following Field’s departure, the focus narrowed to domestic products, specifically robots capable of performing household chores and facilitating video communication.

Despite these ambitious plans, the realization of Apple’s home robots remains uncertain. The challenges of engineering and integrating advanced AI into consumer-friendly products are significant. Nonetheless, Apple’s continued investment in this area indicates a long-term commitment to exploring the potential of home robotics.

Apple is not alone in its quest to bring robotics into the home. Competitors like Amazon have already launched products like the Astro robot, which can perform basic tasks such as delivering snacks and patrolling the house. Additionally, products like the Roomba autonomous vacuum cleaner have set a precedent for consumer acceptance of robotic helpers.

The key to Apple’s potential success in this market lies in its ability to leverage its robust ecosystem of devices and services. By integrating home robots with existing products like the iPhone, iPad, and Apple Watch, Apple could create a seamless and highly functional smart home environment.

Apple has been secretly building home robots, report says

Mistral, a French AI startup, is now offering customization options for its AI models, including a paid version, to help developers and businesses tailor the models for specific tasks. 

Mistral has launched a software development SDK called [Mistral-Finetune](https://github.com/mistralai/mistral-finetune), which enables fine-tuning on workstations, servers, and small datacenter nodes. According to Mistral, fine-tuning a model like Mistral 7B on a dataset like UltraChat, which comprises 1.4 million dialogs, takes approximately half an hour using eight H100 GPUs.

For a more streamlined experience, Mistral provides fine-tuning services through its API, currently compatible with Mistral Small and Mistral 7B, with plans to support additional models soon. Furthermore, Mistral is introducing custom training services, allowing select customers to tailor any Mistral model to their specific needs using their own data. This approach enables businesses to create highly specialized models for their unique domains.

As Mistral seeks to expand its revenue amidst increasing competition in the generative AI sector, it is reportedly seeking a $600 million investment at a $6 billion valuation from investors like DST, General Catalyst, and Lightspeed Venture Partners. Despite releasing multiple models, including a code-generating model, and introducing paid APIs since its first generative model in September 2023, Mistral has not disclosed its user base or revenue figures.

Mistral Unveils AI Model Fine-Tuning and SDK for Developers and Enterprises

on June 6 at 7:20 Texas Time [SpaceX launched Starship](https://x.com/SpaceX/status/1798699175573700929) from the Starbase facility in Boca Chica, Texas. After a few hours of launch, SpaceX confirmed in a [tweet](https://x.com/SpaceX/status/1798715759193096245) we successfully launched the Starship.

[![Watching SpaceX starship lift off](https://firebasestorage.googleapis.com/v0/b/saks-test.appspot.com/o/uploads%2F3pSAcKD69pYhIdjDD4hooYQZIx13%2F20240606195528734.jpeg?alt=media&token=335a85b8-6f6d-42b7-afa9-f7864b4e9fb2)](https://x.com/SpaceX/status/1798699175573700929 "Liftoff of Starship!")
source: SpaceX

It was the fourth flight of the Starship, fully loaded with more than 4500 metric tons (10 million pounds) of fuel, it included the Super Heavy rocket designed to be fully reusable and capable of reaching deep-space destinations like the moon, Mars, or beyond.

[![Watching SpaceX super heavy splashed down in mexico](https://firebasestorage.googleapis.com/v0/b/saks-test.appspot.com/o/uploads%2F3pSAcKD69pYhIdjDD4hooYQZIx13%2F20240606195737601.jpeg?alt=media&token=30d3bcd0-4690-4e70-ae97-9fa39d185344)](https://x.com/SpaceX/status/1798701489097183286 "Falcon heavy splashed down in mexico")
source: SpaceX

after liftoff Super Heavy booster was splashed down in the Gulf of Mexico, which is the most powerful and reusable booster, stands 71 m  tall and contains 33 Raptor engines that produce a combined thrust of 16,700,000 lbs at full power.

SpaceX successfully launched Starship Flight 4, It's confirmed

The company has recently issued a critical advisory for users of its $700 AI Pin. Customers are being urged to discontinue using the Charge Case accessory that came with the device due to potential fire safety risks. This announcement was made via email to customers, with [The Verge](https://www.theverge.com/2024/6/5/24172377/humane-ai-pin-battery-case-issue-warning) sharing the details.

Humane stated that this precautionary measure follows a report of a charging issue. A thorough investigation revealed a quality problem with the battery cell supplied by a third-party vendor, leading Humane to cease working with this supplier. The issue appears to be confined to the Charge Case, and other accessories like the AI Pin itself, Battery Boosters, and Charge Pad remain unaffected. As a temporary solution, Humane is seeking a new battery vendor.

Despite the seriousness of this issue, Humane is not issuing a recall or providing refunds. Instead, customers will receive two months of the Humane subscription for free as compensation. The AI Pin, launched in November, was designed to be a standalone AI-driven device, featuring a magnetic battery system that attaches to clothing, a laser ink display for projecting information onto the user's hand, and a built-in camera. However, the device has been met with significant criticism for its poor performance.

Initially unveiled at CES 2024 in January, the Humane AI Pin was touted as a multifunctional wearable device intended to act as a "second brain and assistant," capable of performing web searches, making calls, sending texts, capturing photos, and taking notes. However, since its release in April, the AI Pin has struggled to meet expectations, garnering predominantly negative reviews.

The recent charging case issue adds to the growing list of challenges for Humane. Reports suggest the company might even be seeking a buyer, indicating deeper underlying problems. The advisory to stop using the Charge Case is just the latest setback for a product that has already faced considerable scrutiny.

This situation underscores the complexities and difficulties of hardware development, especially for new entrants in the market. With nascent categories like AI-driven wearables, the challenges are even more pronounced. The future for Humane remains uncertain, and it will be intriguing to see how the company navigates these turbulent times and what strategic decisions it will make moving forward.

Humane Tells Users to Stop Using Ai Pin Case Over Fire Risk

On Friday night, the dearMoon project—a groundbreaking initiative to send Japanese billionaire Yusaku Maezawa and 10 other individuals on a circumlunar flight aboard SpaceX's Starship—was abruptly canceled. This decision marks a significant setback in the burgeoning field of space tourism, leaving many to wonder about the future of private space exploration.

The announcement of the dearMoon project's cancellation was made on the project's official account on the social media platform X. "It is unfortunate to be announcing that 'dearMoon', the first private circumlunar flight project, will be cancelled," the statement read. "We thank everyone who has supported us and apologize to those who have looked forward to this project."

Shortly after this announcement, Yusaku Maezawa, the project's financial backer and crew leader, provided further insight into the decision. Maezawa explained that when he agreed to the mission in 2018, it was with the expectation that the dearMoon mission would launch by the end of 2023. However, ongoing developmental delays with SpaceX's Starship made it impossible to guarantee a launch date.

#### Reasons Behind the Cancellation

The dearMoon mission was intended to be the first human spaceflight aboard Starship, which would launch from Earth, orbit the Moon, and return. Despite initial optimism, the project faced numerous delays. When the mission was first announced in 2018, both Maezawa and SpaceX CEO Elon Musk were optimistic about the timeline. However, as with many large-scale space projects, unforeseen technical challenges emerged.

SpaceX's Starship program has experienced significant developmental delays. The first test flight did not occur until April 2023, and subsequent tests have continued to highlight the technical challenges of developing such a sophisticated spacecraft. The mission faced additional delays due to the re-prioritization of SpaceX's goals, particularly after NASA selected Starship as the lunar lander for its Artemis Program. This decision injected $2.9 billion into the project and shifted SpaceX's focus towards supporting NASA's ambitious lunar missions.

Since the project's announcement in 2018, Maezawa's net worth has notably declined. However, he has already had the opportunity to experience space travel. In December 2021, he spent 12 days on the International Space Station, traveling there aboard a Russian Soyuz spacecraft. This journey might have impacted his decision to cancel the dearMoon project, as it allowed him to achieve his personal dream of going to space.

The cancellation also affected the eight artists and creators selected to join Maezawa on the dearMoon mission. These individuals, including YouTube creator Tim Dodd and DJ Steve Aoki, had been eagerly preparing for the journey. Dodd expressed his disappointment on X, stating, "I'm extremely disappointed, having dreamt about this mission since I first heard about it in 2018 and even more for the last 3 years since the selection process started."

### Future of SpaceX's Starship

With the dearMoon project canceled, SpaceX can now focus on its immediate priorities: becoming operational, deploying larger Starlink satellites, and meeting NASA's requirements for the Artemis Program. The company's next major milestone will be the fourth test flight of Starship, which is expected to launch as early as June 5, 2024. This test is crucial for demonstrating the technologies needed for in-space refueling and other critical aspects of NASA's lunar missions.

Looking ahead, the next private mission aboard Starship will likely involve another billionaire, Jared Isaacman, who has already flown on SpaceX's Crew Dragon. Isaacman plans at least two more flights before embarking on a pioneering Starship mission.