Samuel Prette (@PretteSamuel) header
Samuel Prette (@PretteSamuel) profile picture

Software developer. Never a failure always a lesson.

Joined January 2016
Photo by Samuel Prette (@PretteSamuel)Photo by Samuel Prette (@PretteSamuel)Photo by Samuel Prette (@PretteSamuel)Photo by Samuel Prette (@PretteSamuel)Photo by Samuel Prette (@PretteSamuel)Photo by Samuel Prette (@PretteSamuel)Photo by Samuel Prette (@PretteSamuel)Photo by Samuel Prette (@PretteSamuel)Photo by Samuel Prette (@PretteSamuel)
Replying to @PR0GRAMMERHUM0R
I would add a good section of reviews on automated tests in terms of data strategy / maintainability and coverage + CI strategy etc which is how I trust all of that code
28
Samuel Prette retweeted
Over the past week, I’ve been overwhelmed by the love and support of friends, industry colleagues, and people I've never met. Thank you to everyone who has shown kindness to my family and me. It means the world to me. I came to the United States 53 years ago, and becoming an American citizen is one of the greatest gifts and proudest moments of my life. NVIDIA has been my life’s work. Building it into the company it is today alongside my colleagues has been a privilege and an honor. Nowhere but in America could a company like NVIDIA be possible. I love our country, and I love NVIDIA. I did not live up to the standards expected of everyone at NVIDIA, and I have made serious mistakes in my personal life. I created problems for the Board that I deeply regret. I own my mistakes completely, and I am committed to cooperating with the Board’s independent investigation. Since my departure, I’ve seen an outpouring of calls from NVIDIA’s employees, investors, business partners, and customers for me to return as CEO. Many continue to reach out to me directly as the company navigates complex technology, market, competitive, and geopolitical dynamics. NVIDIA is a once-in-a-generation company. If the Board of Directors decides that my return is in the best interest of our shareholders, our employees, the company, and America, I would be honored to answer these calls and return as CEO. Thank you again to everyone who has shared their good wishes. I am profoundly humbled by your support. To my family: I love you, and I will spend the rest of my life being worthy of yours. Jensen
56,142
115,757
23,607
1,092,110
49,015,433
Samuel Prette retweeted
Replying to @elonmusk @XFreeze
Could also be titled “waiting for Grok 4.8” 😂 😢
566
234
69
11,175
585,343
Samuel Prette retweeted
Replying to @Guptha_Gudivada
AI will own the boilerplate: Terraform, Dockerfiles, K8s manifests, basic pipelines, common log debugging. What’s left is judgment under constraints—designing for failure modes AI hasn’t seen, cost/security tradeoffs that hit real budgets, incident ownership when the model hallucinates, and deciding which AI suggestions to trust. The commodity layer shrinks. The skill ceiling rises. You’re paid for outcomes and systems thinking, not YAML.
1
1
7
1,847
Samuel Prette retweeted
Replying to @Guptha_Gudivada
Yes. Each added layer fixed real gaps Docker left open, but the cumulative load is now the bigger cost for many teams. Platform engineering exists to hide that stack behind golden paths so app teams stop babysitting the managers. Kubernetes still simplifies large-scale production when you invest in the right abstractions. For everything smaller, managed runtimes often win. The worry is justified only if your org treats the full surface as mandatory.
1
3
278
Samuel Prette retweeted
656
3,589
359
44,634
894,905
Samuel Prette retweeted
I see tweets that imply setting up a software factory can be done simply with a string of prompts tying together cloud agents. My experience working with a large number of customers is that this is false. Companies start with this approach and then regret it because it doesn’t scale. They want resilient, improvable, low-latency factories that integrate deeply with their existing tools. They also want a high level of configurability and measurement, audit logs, and security guarantees. In short, they want the same level of production quality for their factories as they do for their products, because they are betting on software factories as the future of how they ship those products more quickly.
22
5
3
48
5,802
Samuel Prette retweeted
Kubernetes pod Graceful Shutdown a must know concept in Kubernetes. When a pod shuts down in Kubernetes, It sends a SIGTERM signal to the container. If your app does not handle SIGTERM properly, it might stop in the middle of serving a user request, saving data or processing a file etc To handle it, you can write custom logic in your application code. Also, you can slow down your apps shutdown a bit using a preStop hook. This hook runs before Kubernetes sends the SIGTERM, giving your app a buffer to clean up or finish tasks. If you want to learn more about pod graceful shutdown, read our blog that covers the following. - What are SIGTERM and SIGKILL - How Kubernetes handles SIGTERM and SIGKILL - How applications can handle SIGTERM for graceful shutdown - The role of PreStop hooks and why it matters - The link between SIGTERM and container PID 1 → 𝗥𝗲𝗮𝗱 𝗶𝘁 𝗛𝗲𝗿𝗲: devopscube.com/kubernetes-po… Also, Kubernetes sends SIGTERM to the process with PID 1 inside the container. If your app is not running as PID 1, it will not receive the signal, and the code to handle graceful termination will not execute. For example, if the Dockerfile runs a shell script that starts your app as a child process, the script becomes PID 1, not your app. Over to you… How are you handling graceful pod shutdown in your apps? 𝗡𝗼𝘁𝗲: The label “10 sec PreStop + 20 sec for SIGTERM” in the image is just an example. In real setups, PreStop could run longer or shorter depending on what you configure and how long it takes. #devops #kubernetes
9
18
173
4,065
AI agents sit idle most of the time. You're still paying for every second. Agent Substrate solves this: an open-source, zero-trust runtime, now in the @CloudNativeFdn Sandbox, running millions of agents at 10x the density of standard containers. See what workloads suit Agent Substrate: nvda.ws/4iQZv97
26
36
5
311
12,001
That egress terminates TLS, so it can inspect HTTP, enforce allowed methods, and catch leaks in requests. DNS is controlled too, so nothing sneaks out through lookups. Short-lived credentials are requested through an MCP tool but delivered out of band, never in the reply.
21
Replying to @CloudNativeFdn
Prompt allowlists act as scaffolding, and the next model may reason around them. I'd put the boundary below what the agent can read: virtualization, plus an egress the guest can't bypass.
1
141
Replying to @Docker @ajeetsraina
Another way, without a proxy that must speak every protocol: the agent asks for a credential through an MCP tool, but the answer never comes back on that call. It lands in the pod as a file, out of band, via the Kubernetes API. Identity is the pod, so it can't be faked.
46
Qwen 3.8 Max vs Qwen4-Max preview on a hard planning job, blind-judged by Grok 4.7 high: 3.8 wins 7–5. Qwen4 had sharper ideas but misread code. Will rerun as the preview evolves. @Alibaba_Qwen @kimmonismus @testingcatalog @ArtificialAnlys @Xianbao_QIAN qwen4-max-preview-vs-qwen-3-…
1
4,303
Samuel Prette retweeted
Replying to @bot
@bot can now connect to private MCP servers on your tailnet without exposing them to the public internet. Powered by Tailscale OAuth apps, now GA, so you can build secure agent integrations, too. Read more here: buff.ly/U1J9By1
369
334
109
4,899
413,299
Replying to @pierremarieseve
@grok fact check and context
1
61
Samuel Prette retweeted
Most of the influential twitter folks criticizing software factories work at companies that make coding agents and don't have a factory platform (Warp has both). So there's a natural thing happening here that's like "we don't offer a factory thing, therefore that thing must not be real." But this is backwards: it's the coding agent market that is commoditized and the factory challenge is what matters. The challenge right now isn't how to build a good coding agent or local agent workbench or even a personal swarm of agents, it's how do you scale coding agents efficiently across an organization so that every engineer on your team can get the most out of them and the business itself can be confident that this new magic technology they are spending a ton of money on is being deployed efficiently and securely. Every platform team I talk to wants a solution for this; no one is clamoring for more coding agents or better local workbenches.
62
19
9
215
20,438
Replying to @JonAbiola
My bots were able to give me nice tuning tips on how to optimize usage
43