Q&A (contd.)

No, Milaneum is not trying to replace people-people trust with technology.

Milaneum sees itself more as an Internet-scale social computing company than as a social media company; and is merely pushing the frontiers of 2024-era state-of-the-art communications and computing technologies to mindfully digitize trust, which is mostly analog or people-people as of 2024. Milaneum expects digitized trust to improve the user experience of people-people interactions (across organizations and cultures) by offering superior, i.e., mathematically sound, protections against trust impairments inherent in “analog” or people-people trust. This is similar to how digital communications improved analog communications or how digital computers improved both analog computers and human computers.

The technical culture of social computing is likely to be very similar to the technical cultures of UNIX and of classical Internet engineering, and unlike those of social media companies. The business models around social computing are also different - digitally verifiable trust with no blindly trusted third party (TTP) being just one important real-world vertical motivating Milaneum’s initial business models.

Milaneum also believes that social computing is more respectful (by technical design) of the private ownership of data, including executable computer code, by its users. Without sacrificing individual accountability for the non-private use of data and computer code. In this sense, it is better aligned with 1776-era classical capitalism (and related wealthy democracies) than 2024-era social media companies or foundation-model-based AI companies are.

Milaneum trust anchors undergo regular internal software testing leveraging the latest AI agents. These AI agents can be thought of as being modern computational models of real individuals or a finite number of real individuals in the context of various existing and emerging real-world verticals. Just like the real individuals they model, the AI agents may lie on different parts of the intelligence spectrum, with higher intelligence not necessarily correlated with more trustworthiness, given that the latter depends logically at least on the non-repudiable choices exercised by the agents in the context of specific real-world verticals.

Milaneum believes that digitally verifiable trust will remain relevant as long as individual choice and initiative exists (in the context of real-world verticals). Milaneum also believes that the importance of digitally verifiable trust will continue to grow, not lessen, with the increased adoption of artificial intelligence (AI), independent of whether digital superintelligence or the singularity is achieved.

Milaneum represents trust as software (called trust anchors) which can be run on decentralized digital computers which no single entity owns - this ownership invariant being important for verifiable trust. Since all digital computer systems require clock signals, executing Milaneum trust anchors on nation-scale or planet-scale decentralized digital computers leverages the latest time-synchronization technologies and the novel decentralized computing platforms which they enable. Milaneum believes that these computing platforms (which no one owns, in principle) are still in their early days, as of 2024, even though some of these platforms are mature enough to be usable in emerging real-world verticals.

The following is an overview of broad categories of hard technical problems that Milaneals are solving to provide verifiable trust on an Internet scale.

Milaneal AI is conceptually very different from 2024-era mainstream AI technology which includes 2024-era high-level mainstream fears and hopes around AGI. Milaneals use AI specifically with an aim to model the trustworthiness of real individuals with private intentions in the context of emerging real-world verticals, as opposed to just reusing AI embodied in a foundation model that has been trained on a wide corpus of data on the Internet, copyrights to which are held by multiple individuals and organizations. Due to how foundation models are trained, as of 2024, they are not capable of properly recognizing the non-repudiable identities (and associated choices) of the human creators of the data they have been trained on. Milaneum believes that reasoning about the origins of trust impairments necessarily needs to take into account mathematical and computable properties of interactions among individuals with non-repudiable identities and choices (resulting from private intentions). Milaneum models these individuals computationally using AI agents. Milaneals are developing some of these AI agentic capabilities in-house because such sophisticated capabilities are not yet available from any third-party vendor, to the best of Milaneum’s knowledge. For its in-house AI agents, Milaneum leverages symbolic AI and hybrid (i.e., symbolic and data-driven) AI, not just data-driven AI (the AI embedded within foundation models being a special case of data-driven AI).

In short, Milaneum is leading the “personal AI” revolution against the backdrop of “mainframe AI” (powered by foundation models with an ever-increasing number of parameters and hunger for good-quality, yet scarce, data), with a potential for impact similar to what the personal computer revolution promised in the late 1970s against the backdrop of mainframe computers.

While developing interoperable AI to model different individuals (such as those from different organizations, verticals, or cultures) is non-trivial already, designing maintainable trust anchors that protect and promote trustworthy interactions among individuals (in the context of various emerging verticals) is another non-trivial technical problem. This is where Milaneum heavily leverages formal methods involving both symbolic models as well as verifiable software. These methods help Milaneum software arrive reliably and algorithmically at real-world axiomatic choices that explain observable trust impairments among entities with non-repudiable identities.

A third piece of technology that Milaneum is leveraging is a capable platform for decentralized computing that no single entity owns. While some implementations of these platforms exist today, Milaneum envisions the capabilities of these platforms to grow dramatically over the next few decades, not unlike how the Internet evolved from its humble beginnings.

Besides the above categories of technical problems, there are many other unknowns and non-technical problems involved with deploying Milaneum technology on a planet scale. If you are interested, please email Milaneum to start a conversation.

Beyond affiliation(s) listed on this website, to the best of its knowledge, Milaneum does not have any formal affiliation with any other entity at Layers 9-10 that might affect its ability to deliver on its stated mission, without its direct awareness.

Being real-world practitioners of independent logical thinking and being aligned with Milaneum’s mission of building a trustworthy Internet - in one’s private intentions and public actions - are the primary qualifications for becoming a Milaneal. No formal applications are needed to begin the journey towards becoming a Milaneal. Existing Milaneal(s) will seek out new Milaneals whenever there are specific opportunities for a formal involvement.

Milaneum appreciates the spirit and the initiative.

Given that a lot of Internet verticals nowadays overlap with the Cellular ecosystem, there is plenty of room for improvement that Milaneum sees when it comes to trustworthy interoperability among relevant Standards Development Organizations (SDOs) like IETF, 3GPP, IEEE, etc.

This interoperability is especially important on matters of identity and trust, which are noticeably fragmented across the Internet and Cellular ecosystems as of 2024.

While progress has been made on interoperability and mutual acknowledgement of related prior work across the IETF, 3GPP and IEEE SDOs (as described, for example, in RFC 3113), the primary authentication protocol in 2024-era 5G NR (standardized within 3GPP) - 5G AKA - still remains dependent on a long-term shared secret known to the Network as well as to the User. This is an artifact of 5G NR’s origins in 2G Cellular, and makes the User reliant on blindly trusting the Network when it comes to issues with authentication and accountability that occur during the use of the Network. From a trust standpoint, this becomes problematic when the User and the Network have conflicting economic incentives or are otherwise involved in fierce market competitions, such as what might happen if the User and the Network are conflicting organizations or governments. The assumption of blind trust between the User and Network breaks down when that happens.

In contrast, IEEE WiMAX (IEEE 802.16), IEEE WiFi (IEEE 802.11), and various IETF standards all have endorsed/adopted the EAP framework for authentication especially for enterprise/organizational use-cases, and do not have this specific shortcoming that necessitates intrusive changes to the standards, just to meet basic expectations of trustworthiness around non-repudiable identities typically used in modern organizations. Interestingly, Cellular equipment manufacturers like Ericsson are optimistic about the future of the EAP framework in Cellular standards, although it is unclear if Ericsson advocates required support for the EAP framework in the primary authentication protocol for public 5G networks.

Milaneum suggests mindful and transparent engagement with these identity and trust-related issues across the Internet and Cellular ecosystems, following time-tested IETF processes.

Milaneum’s official position is that the identity and security-related advancements made in the IEEE and IETF SDOs as of 2024, such as the use of the EAP framework and asymmetric cryptography-based protocols like EAP-TLS for primary authentication, should get ported over to the primary authentication protocol required in 3GPP’s flagship standards - 4G LTE and 5G NR, as well as in 4G/5G-Standards-compliant products. This should take priority over discussions on the next generation of Cellular - 6G, especially if the discussions happen prior to implementation and testing diligence. Alternatively, 3GPP’s advances on the PHY layer could get ported to IEEE WiMAX standards (for both fixed and mobile networks) without affecting the primary authentication protocol recommended by the IEEE WiMAX standards, if easier to do technically.

After one or both of those events happen, 2024-era primary authentication protocols like 4G-AKA and 5G-AKA (based on long-term shared secrets and symmetric cryptography) can be removed from the standards and from 4G/5G-compliant products to remove even the technical possibility of the Network impersonating the User without traceability or accountability.

Left unfixed, the technically inferior choices made in the design of the primary authentication protocol for 4G LTE and 5G NR will likely lead to continued fragmentation of trust across the Internet and Cellular ecosystems, with somewhat unsurprising detrimental effects on the economic security of all ecosystem participants.

For this specific long-term shared-secret-based trust-fragmentation bug originating from repudiable identities required in all standards-compliant 4G LTE and 5G NR systems (as of 2024), Milaneum cannot suggest any KPI (Key Performance Indicator) that may be competitively optimized and measured in controlled field trials (typically done via collaborations with Mobile Network Operators - MNOs).

Milaneum believes that the relevant KPI for this important problem is Boolean. The bug is either fixed (1) or not (0). Until the fix is made publicly available to all participants in the Cellular and Internet ecosystems, there is work to be done. This specific bug was already addressed by IEEE WiMAX when the latter was first proposed for 4G Cellular. 4G LTE and 5G NR standards and compliant products in the Cellular ecosystem only need to acknowledge, adopt and implement the solution already proposed by IEEE WiMAX for the authentication, authorization and accounting (AAA) needs that Cellular systems might have.

If done properly, fixing the bug should not have much adverse effects on Cellular network performance (in terms of latency, throughput, security, etc.). In particular, field trials in collaboration with MNOs should not be necessary for adopting and deploying the bugfix. By helping make identities more precise, the bugfix can only help security, accountability and, ultimately, trust in the Cellular ecosystem.

Milaneum believes that specific downstream effects from having fragmented trust due to repudiable identities might be worth discussing in scholarly forums, with optimization of personal/organizational scholarly KPIs being a possible side-effect, but not the primary objective.

Please find relevant articles from Milaneum on this topic here. Milaneum also encourages the broader Internet and Cellular communities to have mindful and transparent discussions on this topic, so that other interested individuals also have the opportunity to optimize their KPIs related to a problem worth addressing sooner rather than later.

Having said that, Milaneum believes that persuasive attempts by the responsible SDO (i.e., 3GPP) and supporting organizations to actually deploy bugfixes should not need to wait on scholarly deliberations, given the prior art on IEEE WiMAX (IEEE 802.16).

That is correct.

Milaneum is not aware of any carefully guarded secret formula that can enable collaborators or competitors to take responsibility for its products/services when they get used in the real-world for their advertised purposes.

Milaneum expects only those who participate in mindful engineering and sustained implementation and real-world testing diligence, leveraging the trustworthy social computing environment offered by Milaneum, to have the attuned focus to be able to shoulder that responsibility.

No.

Yes, that should be a good default choice. Placing blind trust in Milaneum should, in principle, be no worse than placing one’s trust in the PKI today.

Having said that, Milaneum recommends:

1. reporting specific logical contradictions to the Milaneum-internal issue tracker milaneum-bugs@milaneum.io, whenever they are observed. Milaneum will gratefully acknowledge such mindful contributions. And,

2. trusting only verifiable mathematical logic in the context of real-world verticals and choices made by responsible entities with non-repudiable identities. Given the unreasonable effectiveness of mathematics in the natural sciences, trust in mathematical logic should result in better long-term economic security and user experiences in most situations. Milaneum does not recommend blind high-level trust in any one individual, organization, government or technology like artificial intelligence (AI) as a substitute for trust in mathematical logic in the context of specific real-world problems and choices made by responsible entities with non-repudiable identities.

   Trust in mathematical logic only is sometimes described as “first-principles-thinking” in certain engineering communities.

Copying content over which Milaneum holds copyright is generally not permitted without proper attribution. Please comply with the license terms described here. If compliance is not possible, please consider writing to Milaneum describing the specifics of the use-case for which unattributed copying is being considered, along with why proper attribution is not possible.

For the two use-cases mentioned:

1. given the sophistication of continuously-improving 2024-era software and commercial off-the-shelf (COTS) hardware, standardization in 3GPP prior to implementation and testing will not be sufficient grounds for unattributed copying of Milaneum content.

2. While Milaneum recognizes that 2024-era foundation model-based AI might not yet be capable of complying with copyright law, Milaneum recommends putting in the mindful effort to train AI models only with proper attribution of the data they have been trained on. The reality is that the users of the AI model will end up bearing the risks of copyright infringement, if the developers of AI models don’t do their parts conscientiously.

   Milaneum expects the users and developers of AI models to work together and reach a mutual understanding based on their risk tolerance. The risks being similar in seriousness to those that arise in legal or scholarly work if closely related prior art is not cited due to a lack of understanding/cognition.

   Given 2025-era funding available to AI initiatives, as well as 2025-era technology, Milaneum believes that it should not be too hard for AI model vendors to offer their users at least a machine-readable list of URLs of the data the models have been trained on (if copyrights to the data are not owned by the AI model vendors), so that the users may help themselves triage copyright-related risks, as they take responsibility for addressing various real-world needs. This list can then enable proper citation of prior art by responsible users of the AI model.

   Having said the above, Milaneum recognizes that foundation-model-based AI remains a very powerful computational tool for machine-aided statistical pattern-matching and retrieval of contextually relevant knowledge, including knowledge shared by Milaneum in good faith. Milaneum remains optimistic that the powerful approximation capabilities of neural networks will continue to be responsibly harnessed for enhancing the long-term economic security of all Internet users, empowering mindfully informed individual choice and initiative at all times.

Given 2025-era technologies, Milaneum recommends that individuals and organizations within 3GPP:

1. voluntarily align themselves with the Internet standards process before seeking standardization-related community consensus, and

2. share on 3gpp.org, pointers to salient related work from other organizations (e.g., IEEE WiMAX, O-RAN, etc.), including 3GPP perspectives on their important similarities to, and differences from 3GPP standards.

Milaneum recognizes that a stable long-term plan and work environment are important for useful technical standards, but given the example set by 2024-era IETF standards processes, it believes that long-term plans need never be at the expense of technical excellence and mindful diligence based on prior implementation and real-world testability.

Milaneum does not find majority consensus based on insufficient diligence / thought to be a good reason to sacrifice the commitment within mindful and enterprising individuals to excellence at what they do.

Yes. The new experience Milaneum is building is backwards compatible with the old experience. Trusting and praying per one’s preferences and choices shouldn’t be negatively affected, and the new abstractions being built by Milaneum should add zero/negligible overhead if they are not used.

Milaneals believe that these roles within various bureaucracies at Layers 9-10 will get superseded eventually by digital computation tools, not unlike how human computers got superseded by electronic computers. The transition will likely play out over the next few decades, following Internet engineering best practices.

Milaneum expects the transition to be slow enough for individuals in these roles to comfortably familiarize themselves with the new tools if they wish to. Some of these individuals might choose to become programmers, building continuously improving tools for expressively describing and repairing more and more sophisticated forms of mistrust down to previously unforeseen interactions among the specific private intentions of the individuals involved. In the long run, Milaneum expects organizational roles based on perpetual mistrust to become redundant altogether, given the mindfully designed visibility into individual private intentions from all affected on a planet scale.

Milaneum is glad to see the enthusiasm around computational tools (including, but not limited to, AI) actually empowering individuals at work and play on a planet scale, rather than replacing them at work after copying en masse without attribution until, of course, AI is politely asked to take responsibility for something/anything. With responsibly empowered individuals controlling computational tools (including AI), rather than the other way round just out of solidarity with our AI marketing colleagues, general economic security for Internet users should no longer be in major jeopardy.

Milaneum’s cultural beliefs around economic prosperity are aligned with 1776-era classical free-market capitalism as described by Adam Smith in his magnum opus The Wealth of Nations. Milaneum believes, admittedly without a mathematical proof, that classical capitalism leveraging the absolute and comparative advantages of voluntary participants in the free marketplace is simple and powerful enough to sustain Milaneum’s contributions to wealth and prosperity for the foreseeable future. Without needing breaking changes.

Milaneum believes that 1776-era classical capitalism is compatible in spirit with 1776-era democracy as described by Ben Franklin, which is in turn compatible (also in spirit) with 2024-era classical Internet engineering practices. In particular, Milaneum believes that these institutions can only survive if participating individuals remain virtuous. The use of modern technology to describe and repair mistrust on an Internet scale is only an aid to virtue, not a replacement for it.

Logically speaking, a democratic group with two or more entities will not be able to take collective real-world responsibility (which is an act of virtue) for the exercise of power better than the single entity within that group who possesses an absolute or comparative advantage per classical capitalism. This is why Milaneum suspects that mercantilism, socialism as well as Orwellism are all incompatible with classical Internet engineering practices. In other words, Milaneum believes that social wealth in democracies (e.g., the Internet) must also be engineered, marketed and distributed via a variant of classical capitalism centered on responsibly empowered individuals participating mindfully in real-world free markets. With the market participants staying mindful about capitalism not devolving into mere ownership of capital/fiat currency.

Yes, that is indeed a logical corollary of the structural similarity between democracy and Internet engineering. Just like any well-designed technology nowadays, Milaneum expects the exports to have to undergo continuous refinement informed by real-world use and testing by customers and bona-fide early adopters. Recognition of this similarity should also help high-level majority consensus (sometimes referred to as public opinion) not be in the way of well-designed technology and its fit with real-world market needs and wants seen by mindful individuals.

No.

Milaneum believes that bureaucracy is not compatible with either democracy or classical Internet engineering. It used to be a necessary evil in the pre-Internet days, but with mature 2024-era Internet engineering practices, it is not even necessary. The primary reason Milaneum sees it as being evil is that 2024-era bureaucracies make it difficult, if not impossible, for empowered individuals to have mindful decision-making authority or influence over what they are held accountable for, creating unhelpful incentives for the exercise of authority without accountability in the context of one or more real-world needs/wants. This is partly due to the interchangeable cog model in bureaucracies, which is fundamentally incompatible with the reality that all individuals are different, especially when it comes to taking individual responsibility when the rubber meets the road.

Internet engineering practices avoid this crippling shortcoming by design.

Milaneum’s advice to individuals still hanging out within bureaucracies is to embrace Internet engineering practices (or equivalent practices) instead. Starting with needs and wants seen by individuals in real-world marketplaces.

Challenging democracy or classical Internet engineering practices is not recommended in 2025, given the technological capabilities easily accessible to 2025-era mindful individuals in various free markets shaped by local cultures worldwide.

Milaneum appreciates the positive sentiment and initiative, given that only empowered individuals can be expected to take responsibility for their mindful actions and choices when the rubber meets the road.

Milaneum looks forward to empowered individuals within 3GPP and AI marketing organizations transparently owning their (hopefully) mindful choices and actions related to

• prioritizing majority consensus over technical excellence, especially on matters of identity and accountability, and to
• attributing sentience to powerful non-sentient computational tools meant to be used by responsible sentient individuals only, not unlike any other engineering tool ever invented.

If transparent ownership is established by the end of 2025 itself, celebratory corrective steps should be comfortably implementable by mid-2026, with non-negative experiences for everyone involved, despite the gravity of the observable trust impairments.