Lightnews — Scholar-powered news

apenwarr

@apenwarr.ca

2.8K followers 1.2K following 1.7K posts

wvdial, bup, sshuttle, netselect, popularity-contest, redo, gfblip, GFiber, and now CEO @Tailscale.com doing WireGuard mesh. Top search result for "epic treatise."

Posts Media Videos Starter Packs

Pinned

apenwarr @apenwarr.ca · Aug 19

So many things are impossible only because someone told you they were impossible and you believed them

3 20 78

apenwarr @apenwarr.ca · 9h

Dude, you can’t leave at the intermission! Godot hasn’t even arrived yet!!

apenwarr @apenwarr.ca · 9h

There’s an AT for that

apenwarr @apenwarr.ca · 9h

Starting branch name

Reposted by apenwarr

BetaKit @betakit.com · 15h

Michael Buhr (C100), Aydin Mirzaee (Fellow AI), Avery Pennarun (Tailscale), and Amanda Arciero (Airudi) break down the challenges and opportunities Canadian AI entrepreneurs face right now, on The BetaKit Podcast.

Listen now:

Are Canadian entrepreneurs willing to let their old companies die? | BetaKit

Leaders discuss the challenges, opportunities, and pressures of AI adoption at ALL IN.

betakit.com

1 1

apenwarr @apenwarr.ca · 13h

On the other hand this suggests a surprisingly simple heuristic for detecting and submerging bullying threads regardless of the topic.

apenwarr @apenwarr.ca · 13h

Yeah. I think we can apply ecosystem diagrams in general: where does energy (money) leave the system, and where does it come in? Incoming, if consumers are paying for all this AI stuff (as it seems they are) then it could be fine. Outgoing, capital investment is less lossy than consumption.

apenwarr @apenwarr.ca · 16h

Sure, I'm in favour of environmental regulations around groundwater in particular. Anything less is nuts.

But, it goes in circles if we make it about datacenters instead of "we are simply not going to use more than X water in this region"... unless datacenters are a significant part of water usage.

apenwarr @apenwarr.ca · 16h

Is there a reference somewhere for what % of the water drain in these endangered systems is from data centers? Because I bet it's less than 2%. Data centers can be built near abundant water sources.

1 1

apenwarr @apenwarr.ca · 16h

Okay, and what happens after it evaporates

1 4

apenwarr @apenwarr.ca · 17h

I'm a thermodynamics sign, not a thermodynamics cop

apenwarr @apenwarr.ca · 17h

And this is entirely separate from the question of what do we mean by "using" water WHERE DO YOU THINK THE WATER GOES

Andy Masley @andymasley.bsky.social · 18h

Every day I find a new way of trying to get across just how ridiculously fake the problem of AI water use is

2 13

Reposted by apenwarr

Andy Masley @andymasley.bsky.social · 18h

Every day I find a new way of trying to get across just how ridiculously fake the problem of AI water use is

12 30 110

apenwarr @apenwarr.ca · 19h

It’s hard to say. If you drew a picture of the entire economy you’d have the same picture of loops of money circulating, just more complicated; that can be healthy.

If you drew a picture of dotcoms reciprocally buying web ads for $millions, it would look like that too and it crashed big time.

1 4

apenwarr @apenwarr.ca · 19h

The ultimate battle: bsky alt-text scolds vs bsky anti-ai scolds!

dame @dame.is · Jun 19

AI/LLMs are a major accessibility technology — being against this technology is to be against advancing human accessibility, and i think there’s a case to be made that it’s borderline ableist

1 13

apenwarr @apenwarr.ca · 1d

The terminology for the different participants in a transaction

The fact that the standard is so bad that two fully compliant implementations can fail to interoperate

2 13

apenwarr @apenwarr.ca · 1d

It’s actually e. Long story

Reposted by apenwarr

𝙃𝙤𝙪𝙨𝙚 𝙤𝙛 𝙇𝙚𝙖𝙫𝙚𝙨 Audiobook Narrator @jefferyharrell.bsky.social · 1d

I posted this last night cause I kind of wanted to bury it. I got cold feet about putting it out there.

embedding-space.github.io/sparse-netwo...

The subject is WHY neural networks work, and I think the answer I offer is kind of interesting. Maybe even a little correct, possibly.

A line chart titled “Accuracy vs. Sparsity (Iterative Magnitude Pruning)” showing model accuracy as weights are pruned. The x-axis represents sparsity from 0% to 100%, and the y-axis represents accuracy from 0% to 100%. A blue line with circular markers shows that accuracy stays around 80% from 0% to roughly 90% sparsity, then drops sharply toward 55% near 100% sparsity. A dashed red horizontal line labeled “80% target” runs across the chart near 80% accuracy, indicating the desired baseline.

11 13 82

Reposted by apenwarr

Justin Garrison @justingarrison.com · 3d

@tailscale.com is still magic

1 18

apenwarr @apenwarr.ca · 1d

Yeah, once you get into a situation with more than one router, some kind of central policy is needed that doesn’t refer to individual ports on individual switches. Kinda weird how long it took people to realize that :)

1 1

apenwarr @apenwarr.ca · 1d

Staycation status

1 21

apenwarr @apenwarr.ca · 1d

Well, we didn’t have ACLs on day 1 of Tailscale but we did add them on day 2 I think, due to overwhelming demand :)

1 4

apenwarr @apenwarr.ca · 1d

Tired: selling shovels for the gold rush

Wired: selling secure turnkey complete mining solutions(tm) for the gold rush

1 20

apenwarr @apenwarr.ca · 1d

I do love San Diego. Stayed at a youth hostel near the beach a couple of times, long ago, and it was awesome. Also I’m told fish tacos exist!

apenwarr @apenwarr.ca · 1d

Ooh, haven't been there. The people in the customer photos on their website look happy though so that’s promising!

apenwarr @apenwarr.ca · 2d

Oh boy. This is a great time to plug one of my favourite extremely niche books, The Infinitr Staircase by Geoffrey Moore!