Rich A Marino rma92

user

Which of these are the fastest for running an LLM, assuming the same VRAM:

Nvidia A100
Nvidia A40
Nvidia L40S
Nvidia GH200

agent

Fastest → slowest for running an LLM (inference/training), ignoring VRAM capacity differences:

agent: ChatGPT 5.2 Auto

user

For some reason, people seem to be funny about using electric screwdrivers / drills / screw guns while working with wiring devices in the US. Is this an accurate assessment? If so, any idea why? I've seen professional electricians use powered drivers - it saves a ton of time and saves stamina!

agent

Short answer: yes, that perception is accurate in the US, and it’s driven more by code interpretation, safety culture, liability, and training norms than by a hard technical prohibition.

Below is the breakdown.

agent: ChatGPT 5.2 Auto

user

How does the current Jeep Wagoneer compare to the Grand Wagoneer? The prices are similar, and I'm looking specifically at renting from National / Enterprise. I routinely see Grand Wagoneers there, and my friend rents them on a regular basis. The "Premium SUV AWD" class lists "Jeep Wagoneer L". Pay particular attention to any notes from Flyertalk.

agent

Here’s a practical, rental-focused breakdown of the Jeep Wagoneer vs Grand Wagoneer — especially relevant if you’re seeing “Premium SUV AWD – Jeep Wagoneer L” in National/Enterprise fleets and your friend often gets Grand Wagoneers. I’ll also include FlyerTalk real-world rental impressions where available.

Agent: ChatGPT 5.2 Pro Thinking

user

Being an experienced software and electrical engineer who also has a ton of experience on the project management, specification, and manual writing side, is it likely that I'm getting better "vibe coding" results than usual considering that I write extremely detailed and thorough specs that break things down into a way that a human could easily complete the various coding tasks?

agent

Yes—very likely.

“Vibe coding” quality is heavily driven by the quality of the problem formulation. If you routinely write specs that a competent human could execute with minimal ambiguity, you’re feeding the model the same kind of structure it needs to produce coherent code: clear intent, constraints, interfaces, and acceptance criteria. Your software + electrical engineering background plus PM/spec/manual-writing experience also tends to reduce two common failure modes: missing requirements and vague definitions.

Set up Podman Compose on Debian

Note: podman-compose uses rootless podman to do the compose. podman compose uses the socket and root daemon.

sudo apt install podman podman-compose podman-docker

systemctl --user enable --now podman.socket
systemctl --user status podman.socket

# confirm it works

user

Do Claude Code or Codex support a way to pass in a very long message from a file? For example, I want to provide the instructions to start building the application. It's several paragraphs, so I want to write it with vim, save it, then tell Claude or Codex to read the file and follow the instructions.

Does this work? Can you find anything online? Ideally try to cover both tools.

Approaches that seem reasonable are:

Writing the file, say, instructions.md or instructions.txt in the directory and telling the agent to read it and build what is described.
Pasting the contents of the file into the interactive prompt.

I'm open to other options. Please try to provide commentary for both tools, as we use both of them at my firm. Thanks for your help, you're the best.

Building a small base for embedded systems on OpenBSD

The following will be removed

clang
relinking sources

Steps

Install OpenBSD normally. I put it in a single partition for processing. Only base.tgz install (and bsd/bsd.mp) are needed.
- If this is a local VM, possibly enable root SSH so that commands can be pasted into terminal more easily.
Boot into the new system with the cd inserted, login as root
Notes:

Alpine ZFS Root

Ideally, follow the steps in "Build a custom virt kernel and ISO with ZFS", and use that install to avoid using unnecessary space and building an unnecessarily large kernel (RAM usage). This produces a system that installs in under 90MB.
Boot from Alpine ISO. Login as root (no password). //* Run setup-alpine to set up network/package repositories, hit ctrl+c when you get to the disk prompt
To set up network with DHCP and the first mirror, run the following, or manually use setup-intefaces -r and setup-apkrepos to set something more specifically. Optionally, use set a root password setup-sshd to set up remote console.

setup-interfaces -ra
setup-apkrepos -1c

echo 'root:9' | chpasswd

	<?xml version="1.0" encoding="utf-8"?>
	<Preset isAutoSaved="true" xmlns="urn:schemas-nliteos-com:pn.v1">
	<Date>11/21/2025 12:42:09</Date>
	<AppInfo>
	<Version>2025.8.10552</Version>
	<Licensed>Yes</Licensed>
	<Protections>Yes</Protections>
	<Host>Windows 10 EnterpriseS 21H2 x64 - 10.0.19044.6575 (en-US)</Host>
	</AppInfo>
	<ImageInfo>

	<?xml version="1.0" encoding="utf-8"?>
	<Preset isAutoSaved="true" xmlns="urn:schemas-nliteos-com:pn.v1">
	<Date>09/12/2022 13:25:18</Date>
	<AppInfo>
	<Version>2.3.8.8890</Version>
	<Licensed>No</Licensed>
	<Protections>Yes</Protections>
	<Host>Windows Server 2022 Standard x64 - 10.0.20348.405 (en-US)</Host>
	</AppInfo>
	<ImageInfo>