Forwarded from Luna's pathway🤗 (Luna)
The hardest part of being a Product Manager isn’t the roadmap or the backlog it’s protecting the team’s energy while holding things together behind the scenes.
That’s why saying no to feature requests, unplanned development, or undocumented work is often misunderstood. It’s not about control or resisting ideas. It’s about protecting clarity. A PM acts as a buffer absorbing pressure, filtering noise, and deciding what should not reach the team yet.
Not every idea needs immediate execution. Not every leadership anxiety should become a sprint task. Unplanned actions and undocumented decisions don’t just add work they drain energy, create confusion, and break momentum. Keeping things simple and understandable is what allows teams to move fast sustainably.
Sometimes everyone has a framework to suggest, a strategy to recommend, or a better way to do things and often, they’re right. But timing matters. Execution isn’t only about knowing what to do; it’s about knowing when it fits. Many times, the direction is already clear it just doesn’t belong right now.
So the no is rarely a rejection. It’s a pause. A not yet. A deliberate choice to hold the line until the work can be done with context, alignment, and care because real product leadership isn’t reactive, it’s intentional.
That’s why saying no to feature requests, unplanned development, or undocumented work is often misunderstood. It’s not about control or resisting ideas. It’s about protecting clarity. A PM acts as a buffer absorbing pressure, filtering noise, and deciding what should not reach the team yet.
Not every idea needs immediate execution. Not every leadership anxiety should become a sprint task. Unplanned actions and undocumented decisions don’t just add work they drain energy, create confusion, and break momentum. Keeping things simple and understandable is what allows teams to move fast sustainably.
Sometimes everyone has a framework to suggest, a strategy to recommend, or a better way to do things and often, they’re right. But timing matters. Execution isn’t only about knowing what to do; it’s about knowing when it fits. Many times, the direction is already clear it just doesn’t belong right now.
So the no is rarely a rejection. It’s a pause. A not yet. A deliberate choice to hold the line until the work can be done with context, alignment, and care because real product leadership isn’t reactive, it’s intentional.
it reminds us of how the Internet really works: smart people think of good things that are crazy enough that they just might work, and then they give them away, over and over, until they slowly take over the world and make things better for everyone.
also, Aaron Swartz, was a beta tester.
[How Markdown took over the world]
If mark_up_ is complicated, then the opposite of that complexity must be… markd_own_. This kind of solution, where it’s so smart it seems obvious in hindsight, is key to Markdown’s success. John worked to make a format that was so simple that anybody could pick it up in a few minutes, and powerful enough that it could help people express pretty much anything that they wanted to include while writing on the internet. At a technical level, it was also easy enough to implement that John could write the code himself to make it work with Movable Type, his publishing tool of choice. (Within days, people had implemented the same feature for most of the other blogging tools of the era; these days, virtually every app that you can type text into ships with Markdown support as a feature on day one.)
also, Aaron Swartz, was a beta tester.
Aaron instantly understood both the potential and the power of Markdown, and was a top-tier beta tester for the technology as it was created. His astute feedback helped finely hone the final product so it was ready for the world, and when Markdown quietly debuted in March of 2004, it was clear that text files around the web were about to get a permanent upgrade.
[How Markdown took over the world]
US Senate unveils crypto market structure draft bill.
The crypto market structure bill that will be the subject of Senate hearings this week.
imo, why this is important:
Stablecoin growth would likely displace bank deposits and reduce bank credit to the real economy, contrary to false claims by the crypto industry.
The risk of significant deposit flight will be exacerbated if stablecoin issuers can indirectly pay interest through affiliates or other third parties.
Crypto lending and borrowing via DeFi platforms present a risk of loss to consumers and could transmit crypto shocks to the broader financial system.
These illicit actors use unhosted and internationally hosted wallets as well as DeFi platforms to evade detection and access the U.S. financial system.
Yield-bearing stablecoins are dollar-pegged tokens that offer returns to holders, typically by sharing interest earned on reserves.
The crypto market structure bill that will be the subject of Senate hearings this week.
imo, why this is important:
Stablecoin growth would likely displace bank deposits and reduce bank credit to the real economy, contrary to false claims by the crypto industry.
The risk of significant deposit flight will be exacerbated if stablecoin issuers can indirectly pay interest through affiliates or other third parties.
Crypto lending and borrowing via DeFi platforms present a risk of loss to consumers and could transmit crypto shocks to the broader financial system.
These illicit actors use unhosted and internationally hosted wallets as well as DeFi platforms to evade detection and access the U.S. financial system.
Yield-bearing stablecoins are dollar-pegged tokens that offer returns to holders, typically by sharing interest earned on reserves.
China's "Are You Dead?" app check in on growing cohort of people living alone.
The app, called Sile Me in Chinese, requires users to “check in” by pressing a button.
https://www.ft.com/content/e92d62d4-26db-40f9-8d4e-a2a84c2e51ec
The app, called Sile Me in Chinese, requires users to “check in” by pressing a button.
https://www.ft.com/content/e92d62d4-26db-40f9-8d4e-a2a84c2e51ec
😁1😢1
Things that used to takes us four month can be done in one afternoon not once but four times. some how it all got normalized quick Jevons Paradox maxxing 🤯
DeepSeek is good at AI research.
One of the things to remember about DeepSeek is that structurally it's closer to Meta than OpenAI or Google - it's an AI lab attached to a business that throws off cash and doesn't really need to monetise the foundational research.
Also, Z.ai China's Zhipu rolls out new AI model trained on Huawei chips
Meanwhile The Senate passed the bipartisan DEFIANCE Act, which would allow victims to sue creators of nonconsensual, sexually-explicit deepfakes, by unanimous consent today.
One of the things to remember about DeepSeek is that structurally it's closer to Meta than OpenAI or Google - it's an AI lab attached to a business that throws off cash and doesn't really need to monetise the foundational research.
Also, Z.ai China's Zhipu rolls out new AI model trained on Huawei chips
Meanwhile The Senate passed the bipartisan DEFIANCE Act, which would allow victims to sue creators of nonconsensual, sexually-explicit deepfakes, by unanimous consent today.
et/acc
The US designates Cambodia's Huione as a money-laundering operation, says it laundered $4B+ from August 2021 to January 2025 for North Korean hackers and others {Selam Gebrekidan/New York Times}
Remember Myanmar? A failed Southeast Asian state of 55 million nestled under China, the country's civil war has had major implications
An inside look at a Myanmar online scam center that opened in 2024 with 3,500 workers from ~30 countries and closed in November 2025 after rebels captured it. Times journalists got a rare look inside one of the compounds where the online fraud industry makes its billions.
https://www.nytimes.com/2026/01/13/world/asia/myanmar-scam-center.html?unlocked_article_code=1.EVA.ttOh.fWepnOt77P8t&smid=url-share
An inside look at a Myanmar online scam center that opened in 2024 with 3,500 workers from ~30 countries and closed in November 2025 after rebels captured it. Times journalists got a rare look inside one of the compounds where the online fraud industry makes its billions.
https://www.nytimes.com/2026/01/13/world/asia/myanmar-scam-center.html?unlocked_article_code=1.EVA.ttOh.fWepnOt77P8t&smid=url-share
NY Times
At This Office Park, Scamming the World Was the Business (Gift Article)
Times journalists got a rare look inside one of the compounds where the online fraud industry makes its billions. Inspirational slogans (“Keep going”) were just the start.
Confer works differently. confer.to
Confer encrypts chat history with keys that never leave your devices. The remaining piece to consider is inference,the moment your prompt reaches an LLM & a response comes back.
Confidential computing
This is the domain of confidential computing.
Confidential computing uses hardware-enforced isolation to run code in a Trusted Execution Environment (TEE). The host machine provides CPU, memory, and power, but cannot access the TEE’s memory or execution state.
LLMs are fundamentally stateless, input in, output out, which makes them ideal for this environment.
When you use an AI service, you’re handing over your thoughts in plaintext. The operator stores them, trains on them, and–inevitably–will monetize them. You get a response; they get everything.
Confer encrypts chat history with keys that never leave your devices. The remaining piece to consider is inference,the moment your prompt reaches an LLM & a response comes back.
Confidential computing
This is the domain of confidential computing.
Confidential computing uses hardware-enforced isolation to run code in a Trusted Execution Environment (TEE). The host machine provides CPU, memory, and power, but cannot access the TEE’s memory or execution state.
LLMs are fundamentally stateless, input in, output out, which makes them ideal for this environment.
For Confer, we run inference inside a confidential VM. Your prompts are encrypted from your device directly into the TEE using Noise Pipes, processed there, and responses are encrypted back. The host never sees plaintext.