IT News Review by Control F5 Software: foldable iPhone – 3D mockup and hints about the "wide fold" format

Foldable iPhone: 3D mockup and hints about the "wide fold" format

Mashable reports on new leaks and hints related to a future foldable iPhone, and the public coverage during this period is linked to the appearance of 3D mockups based on unofficial CADs.

Tom's Guide describes a 3D printable physical model that suggests a wider format, with a 4:3 aspect ratio when opened, distinguishing it from many existing foldables. Screen sizes are also mentioned in rumors and the recurring discussion about a display with a reduced notch.

For the software ecosystem, such a format would push applications towards layouts closer to tablet use in mobile, impacting responsive UI, multitasking, and productivity flows.

Although the information remains at the level of leaks and mockups, the signal is that the market is preparing for a convergence of phone-tablet, and developers will need to treat foldables as a distinct category of screens.

AI in schools: from bans to "how to use it correctly"

Some schools and teachers are trying to move from the reflex reaction of blocking ChatGPT to a more pragmatic approach: training and usage rules for students and teachers, against the backdrop of the fact that AI tools are already being used in real ways.

Instead of treating AI merely as a "shortcut for homework," schools are discussing scenarios in which it can support the teaching process, but also where it can produce unwanted effects (for example, incorrect answers, copying, dependency on assistance).

The technological stakes involve governance and process design: what types of activities remain "AI-free," how originality is verified, what data reaches external platforms, and how the limitations of generative systems are explained in accessible language.

Implementation details vary from one institution to another, but the suggested trend is clear: the discussion is shifting from "does AI exist or not" to "what policies, evaluations, and tools do we have to use it without compromising educational objectives."

AI enters the efficiency phase: cost, energy, and infrastructure become differentiators

The next wave of competition in AI will not just be about larger models, but about efficiency: lower cost per training and inference, energy consumption, predictable operation in data centers.

In this context, former executives and investors from Big Tech emphasize that "whoever wins" will increasingly be linked to infrastructure optimizations: hardware, networking, scheduling, compression, distillation, and mechanisms that reduce compute requirements without compromising quality.

The impact for software and business is direct: AI budget planning is shifting from the experimental area to capex and opex, and the pressure falls on teams that can deliver useful systems with controlled costs, including through simpler architectures and better observability.

Meanwhile, the issue of privacy remains in the background of this race: the increasing volumes of data and the integration of AI into mass-market products put more stress on consent, retention, and control.

Europe between AI competition and climate goals

Europe is at a decision point: either accelerate investments in AI and infrastructure, or risk falling behind the US and Asia, in a context where data centers are large energy consumers.

In the background, the tension arises between two sets of priorities: technological competitiveness and decarbonization. For many companies, this translates into concrete decisions about where to build compute capacity, how to secure energy, and what compromises to accept between speed and compliance.

The implication for the industry is that the "AI strategy" can no longer be separated from the "infrastructure strategy": energy, location, cooling, networking, partnerships, and regulation. In Europe, these variables can matter as much as talent or data.

The theme aligns with the broader trend of increasing electricity demand and public pressure on compute projects, as AI infrastructure becomes visible and politically sensitive.

OpenAI: AI agent browsers will remain exposed to prompt injection

OpenAI warns that prompt injection, attacks in which an AI agent is manipulated to follow malicious instructions hidden in web pages or emails, is a risk that will not completely disappear for agent-based browsers.

OpenAI explains that as "agent mode" expands the attack surface, securing it becomes a permanent race: filtering, isolation, execution policies, tool-level protections, and guardrails for sensitive actions.

For product and security teams, the implication is practical: any agent-based browsing integration must be treated as a system with implicit adversarial input. This means threat modeling, telemetry, granular permissions, and rollback mechanisms.

The central message remains sober: the open web is full of untrustworthy content, and agents that "do things" on behalf of the user increase operational risk if not designed with control and verification in mind.

ChatGPT "Year in Review": annual recap with statistics and usage profile

OpenAI has introduced an annual recap feature for ChatGPT, similar to the "Wrapped" in other applications: the user sees statistics about activity and an AI-generated image that reflects the themes from the conversation.

Along with the number of messages, the recap includes dominant themes, a description of the chat style, the day with the highest activity, and a "archetype" that categorizes the usage, along with a personalized "award."

The feature is available only for users who have enabled the permission for ChatGPT to refer to previous conversations and preferences, linking it directly to privacy and personalization settings.

As a product signal, it is another step towards "AI as a personal system" and towards retention through engagement mechanisms, not just point utility.

Firefox promises a "kill switch" for all AI features, after community backlash

Mozilla has reacted to criticism that arose after statements about the AI direction of Firefox and promised an option to completely disable all AI features in the browser.

The company communicated that users will have clear control to stop AI, and the timeline mentioned for delivery is early 2026, amid concerns about privacy, control, and "opt-out" versus "opt-in."

From a software product perspective, the case is relevant as it shows an increasingly explicit expectation: AI features must be configurable, explainable, and reversible, especially in a browser, where the data surface is enormous.

The signal is that the adoption of AI in mainstream products is no longer just a roadmap issue, but also one of trust, communication, and user control.

Data centers have become a public subject: protests, costs, and politics

This year, data centers have moved from "invisible backend" to a subject of public debate, with protests and civic groups contesting new projects, against the backdrop of massive AI infrastructure expansion.

For the industry, this means that compute plans must also be thought of as social acceptance projects: communication, transparency, local partnerships, justifying consumption, and economic benefits.

The conclusion is that "compute" becomes critical infrastructure, and community reactions can directly influence the pace of delivery of the capacity needed for AI models and products.

Tech companies' debt hits a record, fueled by investments in AI

Debt in the tech sector has risen sharply amid investments in AI. Reuters data shows that by 2025, tech companies will have issued bonds globally amounting to hundreds of billions of dollars, a record for the analyzed period.

The context is that, although large companies traditionally have a lot of cash, the pace of hardware renewal and short obsolescence cycles in AI push budgets towards continuous investments, making debt a more frequent tool than in the past.

The practical implication for business is that "AI at scale" becomes a financial structure decision, not just an R&D one. The cost of capital, revenue predictability, and operational efficiency in data centers are becoming increasingly important.

In a scenario where AI ROI is delayed, pressure on margins and project prioritization may increase, which will be reflected in product selection and execution discipline.

A new file transfer service, built by a WeTransfer co-founder

One of the co-founders of WeTransfer, Nalden, publicly criticizes the company's direction after its acquisition by Bending Spoons and is building a new file transfer service.

The context is familiar in digital products: after acquisitions and restructurings, part of the public perceives product or positioning changes as a departure from initial values, and founders may choose to return with alternatives.

For users and IT teams, the stakes are related to trust and predictability: file transfer means volume, confidentiality, retention, and integration into workflows, and policy or UX changes can generate migrations.

In market terms, the story suggests that "file transfer" remains a competitive space, where differentiation is made through simplicity, control, and a clear relationship with user data.

New York: warning labels for features considered "addictive" on social media

The governor of New York, Kathy Hochul, has signed a law requiring platforms to display warnings for young users before exposure to features such as autoplay and infinite scroll.

The text targets platforms defined by the presence of "addictive" feed mechanisms, push notifications, infinite scroll, and other elements designed to extend time spent in the app, with the possibility of exceptions set by the attorney general.

For the industry, it is a signal of regulation focused on product design, not just content: interface and engagement mechanisms become subjects of compliance.

In practice, this may mean UX changes, age segmentation, auditing of features, and a new layer of "risk management" for teams operating products with algorithmic feeds.

Nvidia and Groq: technology licensing and recruiting the leadership team

Nvidia has concluded a non-exclusive licensing agreement with Groq, a competitor in the AI chip space, and will hire Groq's founder, Jonathan Ross, along with other executives and employees.

At the same time, the difference between the narrative of "acquisition" appearing in the press and Nvidia's positioning, which states that it is not an acquisition of the company, but a transaction on assets and a licensing collaboration, can be noted.

For the AI hardware market, the signal is consolidation: know-how in acceleration and alternative architectures becomes valuable enough to be absorbed through licenses, talent, and asset transactions.

For software teams, the effect is seen indirectly in the availability and diversity of inference platforms: compatibility, toolchains, costs, and deployment options.

Italy asks Meta to suspend the policy blocking rival AI chatbots in WhatsApp

The Italian competition authority (AGCM) has ordered Meta to suspend a policy that prevents companies from using WhatsApp Business tools to offer their own AI chatbots in the app.

The decision comes amid an investigation into possible abuse of dominant position, in the context where Meta is integrating its own chatbot, Meta AI, into WhatsApp.

The stakes involve platform governance: who is allowed to build AI services within a messaging platform and under what conditions, especially when the platform also has its own offering.

For the software ecosystem, the case may influence how third-party providers design distribution and integration, but also how regulators view AI as a competitive layer in existing products.

Alexa+ expands through integrations with commercial and local search services

Amazon is adding new integrations for Alexa+, its AI assistant, so that it can work with Angi, Expedia, Square, and Yelp, starting in 2026.

Through these integrations, the assistant is moving closer to transactional use cases, where the value is not just conversational but related to search, selection, booking, and actions in third-party services.

The implication for business and IT is that AI assistants are moving towards an orchestration model, with connectors and partnerships, raising requirements for security, permissions, observability, and error control.

In the medium term, differentiation between assistants will increasingly depend on the ecosystem, not just on how well the model responds.

Postal and banking services in France affected by a suspected DDoS attack

La Poste, the national postal operator of France, announced a major network outage that affected its IT systems, in the context of a suspected DDoS attack.

DDoS attacks aim to block services by overwhelming the infrastructure with traffic, and the impact in such cases is quickly seen in the availability of applications for the public and in internal operations.

For organizations, the episode is a reminder of the dependence on digital infrastructure in essential services and the need for mitigation plans: scrubbing, rate limiting, redundancy, and communication procedures.

From a risk perspective, the combination of critical services and relatively "cheap" attacks remains one of the constant pressures on public infrastructure.

SoftBank in advanced talks for an acquisition in the data center investment space

SoftBank is in advanced talks to buy DigitalBridge, a firm investing in digital infrastructure, according to a Bloomberg report.

If the deal materializes, it would be a move that strengthens SoftBank's focus on AI infrastructure, in a period where data centers, fiber, and edge are strategic assets.

For the industry, the signal is that the market is maturing: we are no longer just talking about "who makes models," but about who controls the assets that deliver compute and connectivity at scale.

In practical terms, consolidation may influence long-term prices, capacity availability, and the pace of expansion in key regions.

China proposes rules for AI systems "similar to humans"

The Chinese internet regulatory authority has published a draft of rules for AIs that interact with humans and display "human-like" traits or behaviors.

The context is the acceleration of the adoption of conversational and agentic systems, where the boundary between "software" and "social interaction" becomes less clear, and the risks include inducing error, manipulation, and confusion about the nature of the system.

The impact for companies is the possible need for compliance: transparency about the fact that the user is interacting with an AI, rules about simulating identity, and requirements regarding acceptable behaviors.

In practice, such rules may influence the design of conversational experiences and how "personas" for assistants and chatbots are built.

ByteDance: plans for massive investments in AI infrastructure

ByteDance's plans for 2026 include increasing spending on AI, and Reuters has previously documented the company's intention to allocate very large sums for AI-focused capex, including data centers and networking equipment.

In the cited reports, a significant part of the investments is related to infrastructure and semiconductors, reflecting the general trend: the competitive advantage in AI depends on access to compute, not just algorithms.

For the market, this means additional pressure on the supply chain and competition for accelerators, but also an increase in internal capacity to train and run models at scale.

For software teams, such investments accelerate the emergence of AI products integrated into existing ecosystems, with direct effects on advertising, recommendations, and creation tools.

Instagram wants to attract teenagers but risks missing the target

A Washington Post report details Meta's efforts to attract more teenagers to Instagram, including through internal initiatives that try to "relate" to the lifestyle of the young audience.

The mentioned example is symbolic: a space set up as a "museum" with references to meeting places and instructions for "teen-style" selfies, which has raised questions about how authentic this type of internal research is.

From a product perspective, the tension is classic: the platform optimizes engagement and retention, but the teenage audience is sensitive to authenticity and the perception of "marketing done by adults."

For business, the theme remains relevant from both a regulatory and safety angle: any push towards the teen segment comes with increased public attention on design, recommendations, and parental control.

Former FTX US executive raises $35 million for a "perpetual futures" exchange

Brett Harrison, former FTX US executive, has raised $35 million for Architect Financial Technologies, which aims to scale a centralized exchange for perpetual contracts on traditional asset classes.

The reputational context is inevitable: the author reminds of Harrison's connection to the FTX episodes and the fact that his name remains associated with a period of intense scrutiny in the industry.

For the financial market and the digital products space, the subject is relevant as it shows the continuation of the "productization" of derivative instruments into a more accessible and faster format, with a focus on trading infrastructure.

At the same time, the question of compliance and risk remains: such products require solid controls, transparency, and operational management, especially after the lessons of recent years in crypto.

Apple fined in Italy for consent rules in App Tracking Transparency

The Verge reports that the Italian competition authority has fined Apple over 98 million euros, considering that the consent rules for tracking imposed on third-party applications are "excessively burdensome."

The case targets App Tracking Transparency (ATT), where third-party applications must obtain consent through two prompts, while Apple applications can obtain permission in one step, which would reduce the consent rate for advertising profiling and affect ad-based business models.

The implication for the iOS ecosystem is twofold: on one hand, Apple positions ATT as a privacy measure; on the other hand, the regulator treats the difference in treatment as potential abuse of dominant position.

For product teams, the case is an example of how UX and policy decisions become grounds for competition and regulation, not just "best practices" for privacy.

Gmail: the possibility to change the address without losing emails and files

Google may allow users to change the Gmail address associated with the account without losing access to existing emails and files, based on information that appeared on a support page in Hindi.

The old address would function as an alias, and authentication could be possible with either the old or new address, but there would be a limitation on creating other new addresses for 12 months.

From a product and digital identity perspective, the change is significant: the email address is a persistent identifier in the Google ecosystem, and its flexibility involves the correct migration of authentication and aliases.

For users and companies, the potential benefit is reducing friction of "new account vs old account," especially for identities created years ago.

Memory market: AI demand pushes prices and may increase device costs

The explosion of AI infrastructure in the cloud and data centers has increased demand for certain types of memory, and the imbalance between supply and demand is starting to show in prices, with possible effects on smartphones, PCs, and consoles.

The context is that RAM is a critical component for device performance, and when manufacturers pay more in the short term, the pressure can propagate into the final cost of products.

In industry terms, this is a signal of "redistribution" of production capacity towards more profitable areas for AI, which can create relative shortages and price volatility in the rest of the market.

The Verge notes separately that consumer hardware manufacturers are also feeling cost increases for DDR5 modules, attributing the phenomenon to a global shortage that may continue into 2026.

Deepfakes become a weapon of social engineering

Deepfakes no longer just mean cloned clips or voices, but attacks on digital identity, with direct risks for trust in authentication and verification.

The article explains that, unlike classic fraud based on stolen data, deepfakes can create new, credible identities, with faces, voices, and documents that pass superficial checks from the very first interaction.

The impact for organizations is in three directions: authentication mechanisms become fragile when relying on "replayable" signals, fraud scales industrially through automated identity generation, and onboarding and KYC processes are put under pressure.

In practical terms, the discussion shifts towards detection, liveness, context signals, and risk policies that do not treat identity as a "file," but as a set of dynamic evidence.

Andrew Ng: AI is impressive, but still limited and will not replace humans anytime soon

In an NBC News interview, also picked up by AOL, he quotes Andrew Ng saying that the difficult part about AI is to understand simultaneously how amazing it is and how limited it remains.

Ng describes AI as an ongoing revolution and anticipates continuous progress, but warns that exaggerations about the rapid replacement of humans do not reflect the technical and operational reality on the ground.

For IT and business professionals, the message is pragmatic: value comes from integrating AI into concrete flows, with validation and accountability, not from expectations of "complete autopilot" in complex organizations.

In the current context, Ng's position serves as a useful counterbalance to hype narratives, especially when investment and reorganization decisions are made under pressure.

Sergey Brin on Google Glass: "I tried to commercialize it too quickly"

Times of India reports on a talk by Sergey Brin at Stanford, where the Google co-founder discussed the failure of Google Glass and the lessons related to the premature launch of a hardware product.

Brin says he treated the launch as a "Steve Jobs" moment, but the product was not mature enough: neither in finish for consumers, nor in cost level, nor in experience maturity.

The implication for product teams is a classic one, but relevant in the AI hardware and wearables era: a spectacular demo does not replace the complete "baking" of the product, and the time spent on robustness, cost, and UX can determine success.

In a market that constantly pushes for "first to market," the Glass example remains a lesson about timing and the difference between prototype and mass product.

Synthesis realized with the help of a monitoring flow provided by Control F5 Software.