wmf

52,098 karma · joined August 19, 2007

Wes Felter, expert in datacenter/cloud networking and containers. Previously worked at IBM Research and a large investment bank.

submissions comments

wmf··on Cheap Desktop 400GbE Switch MikroTik CRS804-4DDQ-HRM Review

They're not connected to the ASIC so anything on those ports is taking the slow path.

wmf··on .name Termination

This is how you fill out an application when you know it's going to be rubber-stamped.

wmf··on .name Termination

Discussion from yesterday: https://news.ycombinator.com/item?id=49516047

wmf··on Ask HN: How to get taken seriously while using AI?

The main thing is to never mention AI. Don't say that you use AI and don't leave obvious evidence. Don't lie, but don't talk about AI unless someone asks.

Also maybe try to find a community that aligns with your values (not lobsters).

wmf··on Discontinuation of third level domain registrations for the .name TLD [pdf]

Instead of registering your.name you would register first.last.name (nobody owned last.name). This was intended to allow more people to have personal domains with fewer conflicts.

Now, years later, they are undoing it and they're even breaking existing domains. ("existing third level domain names will be terminated") This kind of breaking change shouldn't be allowed.

wmf··on How accurate have Ed Zitron's AI skeptic predictions been?

https://danluu.com/ballmer/

wmf··on How accurate have Ed Zitron's AI skeptic predictions been?

Speaking of wrong predictions... the entire US military budget would have to be spent on OpenAI/Anthropic to keep the bubble from bursting.

wmf··on How accurate have Ed Zitron's AI skeptic predictions been?

...get any media presence, because that demands being in front of people quite often, and basically nobody can be insightful and well researched in all topics of the day

What? Zitron is getting a lot of media attention by repeating the same takes hundreds of times. He's not a general pundit.

wmf··on Understanding ChatGPT Work

It seems the same to me; even the naming is extremely similar.

OpenAI Codex = Claude Code

OpenAI Work = Claude Cowork

wmf··on Internet centralization and the original sin of NAT

I don't think any of us could know how most users use it.

wmf··on Internet centralization and the original sin of NAT

Can't every chat app do this now, including SMS/MMS?

wmf··on IBM Unveils Next Generation Dual-Architecture Processor for IBM Z and LinuxONE

I don't think z has been benchmarked in 20 years.

wmf··on IBM Unveils Next Generation Dual-Architecture Processor for IBM Z and LinuxONE

It starts in z mode and can only run ARM code in VMs.

start as a regular x86 but then be able to execute code of other architectures

https://en.wikipedia.org/wiki/Alternate_Instruction_Set

wmf··on IBM Unveils Next Generation Dual-Architecture Processor for IBM Z and LinuxONE

The previous A64fx was just for HPC. Monaka looks more general purpose so maybe they will sell Monaka servers.

wmf··on Understanding ChatGPT Work

I suspect it's mostly marketing. People think "I don't want code so I won't use Codex" so they get the same thing but labeled Work.

wmf··on Internet centralization and the original sin of NAT

I'm having trouble decoding this but Tailscale is a VPN that can traverse NAT when necessary. Arguably Tailscale wouldn't be needed if NAT didn't exist but considering the timeline I'm not going to declare a conflict of interest.

wmf··on Meta's Push to Put Robots to Work in Data Centers

They have BMCs in each server but if the BMC doesn't respond they are probably using humans to physically unplug the sled.

wmf··on Meta's Push to Put Robots to Work in Data Centers

Meta doesn't use enterprise equipment (for example the "PDU" is a passive DC bus bar); they've chosen to make a tradeoff between capex and service cost.

wmf··on IBM Unveils Next Generation Dual-Architecture Processor for IBM Z and LinuxONE

It's not entirely clear what you're asking but... AGI, Vera, X2, Spark.

wmf··on IBM Unveils Next Generation Dual-Architecture Processor for IBM Z and LinuxONE

That may be true but the dozens of x86 servers are cheaper and easier to use.

wmf··on IBM Unveils Next Generation Dual-Architecture Processor for IBM Z and LinuxONE

Because ARM is the future.

wmf··on IBM Unveils Next Generation Dual-Architecture Processor for IBM Z and LinuxONE

I think Nintendo used Tegra because it was super cheap. AMD has access to the same ARM cores as everyone else (see Seattle and Sound Waves) but a semi-custom chip would have been more expensive than an overstocked Tegra.

wmf··on IBM Unveils Next Generation Dual-Architecture Processor for IBM Z and LinuxONE

The existing 1,000 customers replace their old mainframes with newer models.

wmf··on IBM Unveils Next Generation Dual-Architecture Processor for IBM Z and LinuxONE

The z mainframe team does good technical work but then the high price cancels out all the value of that work. I'm not sure if that counts as innovation or not.

wmf··on VMs won't contain cyber-capable agents

More like QEMU won't contain agents.

wmf··on New Mac Studio with M5 Max and M5 Ultra

AFAIK these "rules" are artificial and there's nothing stopping Apple from having twice as much RAM. They just don't want to.

wmf··on Apple introduces M6 and M5 Ultra

That model would be larger and more expensive than the entire M7 chip.

wmf··on Musk's SpaceX to build $100B launch facility in Louisiana

SpaceX is going all in on Starship.

wmf··on OpenAI Jalapeño: Better than Nvidia Blackwell

Existing CPUs have been extremely optimized by ~6 competing, well-funded teams. I expect AI to accelerate things somewhat but it's not clear that there is any low-hanging fruit available for AI to find.

wmf··on OpenAI Jalapeño: Better than Nvidia Blackwell

Every company is designing their own chips so the dominant players will be one level down: Broadcom, TSMC, Hynix/Samsung/Micron, etc.

← PreviousPage 5 of 34Next →