By Lance Maurer, CEO

Engineering is supposed to be fun.

Developing solutions for tough technical problems with customers who have become friends while using collective brain power and modern technology is a rich and satisfying professional experience.

Yet, it often feels as though our technological gains are invariably met with a constant avalanche of change in the “name of progress”. This happens on both sides of the relationship, and often manifests as consolidation, budget cuts, “organizational shifts,” new technology and consumer trends, and an increasing pressure from stakeholders for teams to deliver more with fewer resources. Painful as it is, the truth is that consolidation is an unrelenting force in our industry—it will persist even after we all end up working for the Borg! This is simply one of many core natures of our universe; like the curious space dust that began clumping together 14 billion years ago, things will continue to merge and shift in strange ways through the eons – every jungle survival story follows a similar pattern.

So if industry chaos is the norm, and jarring technology evolution is merciless, how do we design technology to retool factories that must operate efficiently through these dynamic changes? 

In my professional Odyssey through two different industries (aerospace and now media tech), I have grown to accept the natural laws of merging chaos as foundational truths – they have become part of the design process itself, necessary for complex machines to ensure longevity and for businesses to endure.  It does not matter if you are the technology vendor or the customer applying it, understanding said ground truth is crucial to how the design and buying process should happen so that both parties thrive.  Investment should be made in technology that is adaptable to change, with an attention to a 5-7 year horizon.  When looking at technology to invest in (not only money, but time and trust), start with aligning these baseline truths – does your business share core technology values with the vendor you are considering?

My view of the “Media Supply Chain” technology core values:

1. Making consistent, professional media is creatively, technically and financially challenging
2. Platforms should be hybrid, scalable, expandable, and affordable to provide maximum flexibility
3. Industry standards provide format direction, but consumption methods (e.g. 9:16) are shifting and tools need to be able to grow outside the current box
4. Output quality matters – artifacts are avoidable
5. A.I. has great potential but should be handled with great care
6. Our industry is in a (relatively) high transformational period, which will likely continue for some time

After core values, the keys to adaptability over time:

Hybrid strategies thrive in times of high change…

Toyota did it right. Their “multi-pathway” strategy—a calculated bet on hybrid engines—has proven to be the definitive answer for the current auto market, as evidenced by their surging sales figures. By prioritizing solutions that can work flexibly on prem and in the cloud (“hybrid”), we can bridge the gap between proven, dependable systems and the cutting edge, without exposing much risk.  This balanced approach to planning and purchasing decisions is vital when nascent and untested new technologies are unleashed into the wild. Because the media supply chain is a complex engineering ecosystem, it requires more than just a leap of faith into the new; it requires an accepted foundation of what already works and alongside an ability to experiment and take advantage of new approaches as well.

AI strategies demand a similar hybridization approach for supply chain technology like image and audio optimization. The media supply chain tech stack should include several “workhorse” foundations of proven technology paths augmented by increasingly validated new tech (like AI) to work toward specific efficiency targets. Unless you have an appetite for risk in your R&D budgets, going “all-in” on AI for more complicated workflows like image optimization is a high-stakes gamble, as the tech is still evolving and companies attempting to create products are coming and going.  Prioritizing iterative process AI over generative models makes sense because determinism offers “set-and-forget” reliability to free you up for the next task; conversely, when AI attempts to be “clever” through generation, it often introduces unreliably generated artifacts into workflows.  This is why it is key to enable the trusted humans here because they know from experience where to look for gaps in these new tools, avoiding potential hazards; this tacit knowledge, or “know how” has not transmuted into the AI models …

Technology is a tool for a great team – prioritize the human process and supplement with great tech to deliver the best culture of technical growth.

Cloud infrastructure serves as a perfect case study for hybrid necessity:

• • The Baseline: Building on-premise (customer-managed infrastructure, private VPC, etc.) systems to handle the bulk of your standard workload remains a logical move given today’s cloud economics.

• • The Burst: Public hyperscalars are an unparalleled tool for “pants-on-fire” surges or short-term projects – it is more expensive, but highly predictable and relatively easy to navigate.  An “overnight delivery” premium would cover its cost.  In addition, a lot of great, experimental media technology calls it home.

• • The Future: While interconnectivity through cloud is undoubtedly the future for LIVE events, a purely software-defined, end-to-end, highly reliable 2110 cloud broadcast solution is still a few beats from total maturity.  That said, compressed formats for lower tiered production works fine today.

By designing your public cloud workflow to handle only some fraction of your peak load, you keep the complexity manageable, and the pricetags predictable. This allows your team to learn the nuances of hyperscalers—navigating their advantages and disadvantages—without over-leveraging your entire operation on an evolving architecture.

Picking vendors who provide financial operation visibility layers to enable realtime “finops” analysis is another major component of the adaptable platform concept.

Intelligent Platforms have Longevity

Amplified by machine learning, an “intelligent platform” is characterized by its capacity for both horizontal and vertical expansion. This growth is achieved by integrating diverse, modern technologies to enhance the capabilities and output derived from a core source data set. Over time, as the platform learns from experience and the repetition of complex problem-solving, it evolves to automate more processes and make non-generative decisions for the user—this is the “intelligence” imbued by its creators.  AI will improve the fluidity of this growth in the years to come.

This constant evolution allows the platform to solve a wider array of problems using the same foundational data. The impetus for this growth often comes from external factors, such as consolidation forcing the adoption of new methodologies, or the natural evolution of processes leading to greater efficiency. Ultimately, it addresses the continuous management pressure to deliver more significant results in less time.

My early aerospace work involved time-consuming thermal analysis using a NASA Excel program, taking months for complex inputs and requiring 2D mapping to visualize results. It was highly prone to human error, but was lightyears beyond hand-calculations done in previous years.  Design relied on 2D AutoCAD files with layers, necessitating dozens of drafters and hours of hand-drawn iterations.

The shift to advanced, intelligent 3D modeling revolutionized our entire process.  There were plenty of options on the market, but we chose the platform that had the best answers to where we thought we needed to go and it paid off. The new consolidated platform integrated 3D creation, fabrication drawings, parts lists, and analysis from a single 3D model dataset. Our small team leveraged this integration to streamline design and analysis (thermal, structural, mass properties), enabling us to outpace larger competitors with a small team.  The technology did not force the company to let anyone go, it made who was there more valuable.

This same principle can apply to the media supply chain as well.

AI as an Enabler

It’s going to all be software soon, and we’ll need help.

Software is poised to dominate media workflows, for both file-based and live content. While dedicated hardware with ASICs will retain a niche and be necessary for a while (supporting the hybrid approach), smart agents will ultimately keep software running efficiently with less human oversight running on infrastructure that can pop into and out of existence to optimize financial efficiency.  AI and Cloud are great enablers to this cause, and are critical for the evolutionary steps.  People (especially subject matter experts) will remain essential for ensuring the system works, is tracking, and maintains the ability to deliver the narrative without errors (in addition to consistently envisioning the next generation of improvements).

Artificial Intelligence will not supplant human ingenuity and imagination. Ingenuity and imagination allow us to find novel solutions beyond the expected norm – which will, hopefully, be the final differentiator between man and machine.  AI, by definition, is confined to “the box”—a reflection of the information it is fed, presenting the most probable outcome as the single best answer. For example, if AI existed in 1540, it would have validated the Earth-centric universe while providing “proof” that only witches can float based on the data available then.  Ingenuity and imagination break through barriers and advance us, with technology as the enabler. AI has a place in the media supply chain, especially for automating tedious tasks. However, the design and management of the entire tool stack should remain a human endeavor. Solutions should be built on intelligent, expandable platforms, utilizing APIs that solve both current and anticipated future problems.

—

Consolidation and technological evolution are relentless and inevitable.  We must design and invest with a thoughtful outlook toward these facts, using technologies that can expand around our existing needs as they change while ensuring today’s problems are solved with precision.  The retooling never really stops, then, much like our own learning.  Good technology grows with us, and empowers the individual to do more with less, while keeping the business healthy.  The stakeholders should appreciate that.

Come see us at NAB April 19-22 in La Vegas at Booth #W1547 to see our PixelStrings PIX workflow AI helper tool, NVIDIA AI upres, and more.