DEV Community: Chandan Kumar

AWS drops NDAs for data center energy use, but the gap remains

Chandan Kumar — Sat, 03 Oct 2026 19:10:38 +0000

AWS no longer requires NDAs for data center electricity contracts. That is the headline from AWS CEO Matt Garman’s recent response to community backlash, as reported by TechCrunch. The change signals a shift, but for developers building on AWS, the real question is whether this changes anything about how you evaluate a cloud provider's environmental footprint.

What actually changed

Garman stated that AWS has stopped using non-disclosure agreements in contracts with local utilities and data center operators. Previously, these NDAs prevented communities and local governments from knowing how much power a specific AWS data center consumed. The move follows years of criticism that cloud providers operate in an energy transparency vacuum.

Why this is still not enough

Dropping NDAs from utility contracts does not mean AWS is publishing per-datacenter energy data. It means the utility is no longer contractually gagged. The data may still be hard to find, unstandardized, or buried in public utility filings that vary by jurisdiction. For a developer who wants to know the carbon intensity of their workload running in us-east-1, the practical path to that number remains unclear.

What practitioners should watch for

If you are evaluating cloud providers on sustainability, look for the following concrete signals:

Per-region or per-availability-zone energy usage reports published on a regular cadence
Third-party audited carbon accounting (not just self-reported offsets)
Real-time or near-real-time grid mix data for the regions you use

AWS’s own customer carbon footprint tool gives you a monthly estimate, but it is an average. For latency-sensitive or batch workloads that can shift to greener hours, averages are not actionable.

The bottom line

This is a step toward transparency, not a finished solution. AWS removed a contractual barrier, but the data itself is still not surfaced in a developer-ready way. Until per-zone energy data is available via API or dashboard, the trust gap remains.

Source: https://techcrunch.com/2026/10/03/amazon-responds-to-data-center-backlash-says-it-no-longer-uses-ndas

Source: https://techcrunch.com/2026/10/03/amazon-responds-to-data-center-backlash-says-it-no-longer-uses-ndas/

Meta open sources the code to build your own Muse AI companion gadgets

Chandan Kumar — Fri, 02 Oct 2026 22:04:24 +0000

Meta just open-sourced the core code that powers its Muse AI agent, and now you can build your own physical gadgets around it.

What you get

Meta released the software stack that runs Muse, its conversational AI agent. The code is available so anyone can load it onto custom hardware. Meta suggests several reference builds: putting Muse on a color E Ink display for showing reminders, or embedding it into an HDMI stick to display on a large screen.

This is different from just calling an API. You get the actual agent runtime, meaning your device can run Muse locally or with minimal cloud dependency, depending on your build.

How to use it

Download the open source code from Meta's repository.
Pick a compatible display (color E Ink is confirmed) or an HDMI output board.
Flash the firmware and configure your agent preferences.
Connect to power and a network — Muse handles the rest.

The HDMI stick approach is particularly interesting for developers. You can plug it into any monitor or TV and have a persistent AI assistant visible at all times, without needing a dedicated screen.

Why this matters for builders

This is a shift from the usual "we ship a device, you buy it" model. By open-sourcing the code, Meta lets you control the hardware, the form factor, and the privacy boundaries. You can build a Muse gadget that only connects to your local network, or one that surfaces specific data from your own services.

For now, the builds focus on display-only interaction — showing reminders and information. But the underlying agent can handle voice and text conversations, so future hardware could include microphones and speakers.

If you've been waiting for an AI companion you can actually own and modify, this is the first major release that makes that possible without reverse-engineering a closed device.

Source: The Verge — Meta open sources code to let you make Muse AI gadgets

Source: https://www.theverge.com/tech/1004330/meta-muse-ai-gadgets-home-link

Why the Pope's AI Art Critique Matters for Your Next Model Choice

Chandan Kumar — Fri, 02 Oct 2026 15:55:14 +0000

Pope Leo XIV’s recent statement on AI-generated art draws a clear line that every ML practitioner should internalize: "There is an ontological difference, even before an aesthetic one, between art and what a machine can generate through statistical calculation based on millions of images created by others." This isn't just a philosophical argument—it's a constraint that should shape how you evaluate and deploy generative models.

The Ontological Gap in Practice

The Pope’s core claim—that algorithms lack "the spark of humanity"—maps directly to a technical limitation in current architectures. When you train a diffusion model or large language model on a dataset of human-created works, you are compressing a distribution of human expression into a statistical approximation. That approximation captures patterns, but it cannot replicate the intentionality, context, or lived experience behind the original work.

For your next project, this means you need to decide: are you building a tool that mimics human output, or one that augments human creativity? If the latter, your architecture choices change. You might favor models designed for interactive refinement (like ControlNet or InstructPix2Pix) over end-to-end generation pipelines. You might also invest in provenance tracking—knowing which training samples influenced a given output—so users can trace the “statistical calculation” back to its human sources.

Benchmarking What Matters

Standard benchmarks like FID or CLIP score measure aesthetic similarity or prompt alignment. They don’t measure originality, intent, or the “spark” the Pope describes. If you’re shipping a creative tool, consider adding a human evaluation rubric that asks: Does this output feel like it was made by someone, or just from someone else’s data? Several startups (e.g., Spawning, Bria) now offer datasets and APIs that filter training data to only include opt-in human works. Using them might not make your model more “human,” but it respects the ontological difference the Pope identifies.

The Pope’s letter is a reminder that the gap between art and algorithmic generation isn’t a bug to be fixed—it’s a feature of the technology itself. Design your systems to respect that gap, and you’ll build tools that creators actually trust.

Source: Pope Leo XIV is not a fan of AI-generated art