In the Fall of 2024, the Pacific Telecom Conference (PTC) took the unusual initiative to schedule a multi-day event in Washington, DC. This was a first and Interglobix was there. While PTC has had a long history of annual meetings in Hawaii, this venue opened up participation to many government agencies and businesses in the DC area that might not have had opportunity to engage on the rapidly evolving global telecommunications infrastructure and the myriad applications that are emerging from this rich substrate. Access to the Internet has been dramatically increased thanks to proliferating Low Earth Orbiting Satellites such as Starlink from SpaceX. Moreover, massive new investments are underway to increase the capacity and resilience of subsea cable systems. Google, for example, is in the midst of a significant architectural restructuring aimed at both of these important aspects.
Rapid Response Computing
What might we foresee for 2025 in the way of increasing infrastructure and application development? The demand for computing power (I still find the term “compute” weird when used as a noun) is growing, not in small measure stimulated by new applications of machine learning, including large language models. Moreover, computing at the edge is similarly growing, in part to meet the demand for low latency computation. Just as Content Distribution Networks took off to deliver content at speed and quickly, there is now demand for rapid response computing cycles. The proliferation of the Internet of Things continues and with it, the need for much improved safety, security and confidentiality.
Reliability, Safety, and Security
Our increasing dependence on the Internet and mobile technology continues unabated and should give everyone pause. The headlines announcing outages (e.g. CrowdStrike) and ransomware attacks are just the tip of the iceberg. Single points of failure and dependencies are creating the potential for widespread cascade failures. All of this increases the need for more attention to systems level analyses, software reliability and security. It should also be apparent that safety and security are not synonymous. One could deliver messages very securely only to have them launch phishing or other kinds of attacks once they are decrypted and interpreted or read.
Accountability
At the Internet Governance Forum taking place in Riyadh literally as I write this essay, many of these topics are getting attention, together with increasing concerns for preserving an open, accessible, affordable, usable, reliable and trustworthy Internet. As governments try to come to grips with online abuse, they struggle to assure that human rights are respected while users are also protected from harm. A theme that is increasingly evident is the importance of accountability in this complex environment. I have written on this topic before and believe it will be a key topic as the Global Digital Compact and the Pact for the Future are elaborated in various UN and other global settings.
Standards Maturity
A personally exciting theme for 2025 is the maturing effort to standardize, test and deploy the basic elements of an Interplanetary Solar System internet (SSI). The effort is well underway in a collaboration among NASA, ESA, JAXA and KARI, facilitated by standards work in the Consultative Committee on Space Data Systems (CCSDS) and the Internet Engineering Task Force (IETF). Other space agencies are moving ahead to provide communications capability for missions to the Moon and Mars and beyond.
The AI Agent
An emerging capability in the rich and rapidly evolving Artificial Intelligence space is the so-called AI Agent. A large language model capable of interacting with humans but also carrying out tasks on behalf of users. There are serious efforts to facilitate inter-agent interactions by way of text-based exchanges. Such efforts are simultaneously exciting and challenging. Humans have enough trouble communicating clearly through language. Interbotic interactions by this means may reveal some misunderstandings leading to unexpected and probably unintended consequences. My guess is that researchers will be looking for ways to reduce ambiguity in these interactions to limit the potential for mishaps. Formalisms such as those found at schema.org and semantic dictionaries may help to contain the potential for hazardous outcomes.
Climate Change
Against all of this, we are collectively recognizing that climate change is an increasing threat and made manifest by fires, increasingly severe storms, coastal damage from rising sea levels among other risks. Researchers are accumulating more data from a variety of terrestrial and space-based sensors. Sharing of access to this data will help us to collectively make better risk predictions. I have been much impressed by the use of machine learning models to make high quality, near-term weather predictions requiring much less computing power than the more traditional Navier-Stokes path to weather system emulation and projection. We are far from exhausting the potential for machine-learning tools and applications and 2025 is bound to surface some exciting and likely unanticipated successes. As readers will be aware, several Nobel Prizes were awarded in 2024 reflecting the remarkable capabilities these tools have already demonstrated.