fnerdman

Subject Matter Expert Heterogeneous Confidential Compute @ Ritual

[fnerdman @fnerdman]

@jakehemmerle Who's the supplier?

[fnerdman @fnerdman]

@S1r1u5_ 'unfortunately' ?

[fnerdman @fnerdman]

@jakehemmerle What's the case? And how are the thermals? Did you go for the arctic 4k fans? I'm cooling 2 170hx's with a single arctic 4k and they go above 80 degrees and start throttling at max fan speed. I found the fan its self to be pretty quiet, but the shroud and vibrations cause noise.

[fnerdman @fnerdman]

@socrates1024 Get married, have kids. You won't wish for open problems any longer.

[fnerdman @fnerdman]

@lukewrightmain @CoreX_Hosting I'll take all 200

[fnerdman @fnerdman]

Three honest options for building on TEEs today:
- Phala/dstack: native web3 x TEE (+GPU) Platform, pushing PoC
- GCP vanilla TDX: most promising long-term, no operator endorsement now
- Azure / Nitro / GCP Confidential Space: solid endorsement today, not expected to evolve 10/10

[fnerdman @fnerdman]

I was once asked whether we can trust a billion dollars to a Web3 TEE project. My answer: "No, but I believe we can get there." Permissionlessness TEE execution is the next big thing we need to solve. Where it stands 🧵

https://t.co/3EhyvodMjQ

[fnerdman @fnerdman]

PoE and PoC unlock endorsement for non-hyperscalers. Intel PoE: any operator signs a Layer 3 endorsement of the DCAP quote - OVH, OpenMetal, colocation. Proof of Cloud: parallel path via tiered registry. L1 (witness) ships but doesn't scale. L2 isn't built yet. /9

[fnerdman @fnerdman]

👀 @gajesh

[thaidn @XorNinja]

Early this week, @brucedang and I had a meeting at Apple Park in Cupertino. While there, we also shared with Apple our latest vulnerability research report: the first public macOS kernel memory corruption exploit on M5 silicon, surviving MIE. It was laser printed, in honor of our hacker friends.

We wanted to report it in person, instead of getting buried in the submission flood that some unfortunate Pwn2Own participants just experienced. Most respected hackers avoid human interaction whenever possible, so this physical strategy may give us a slight edge in the eternal race for five minutes of fame and glory on Twitter.

This is the story of the exploit and our field trip. Full technical details will be shared after Apple fixes the vulnerabilities and attack path. Hopefully it won’t take our beloved company too long. We only budgeted one year of domain registration fees for this attack.

This is our strongest research yet, led by @justdionysus, @blacktop__ and @brucedang. It is really dope.

Full story: https://t.co/A4w6cJOAFa

[Ritual Foundation @ritualfnd]

Ritual testnet is now live. The first chain where an agent's existence doesn't end when its creators do.

[fnerdman @fnerdman]

@bgmshana Pro tip: don't use the same payment method

[fnerdman @fnerdman]

@mteamisloading Attention degradation:
https://t.co/6wKnE5swNg

[fnerdman @fnerdman]

Noticed your AI agent deteriorating during marathon vibe sessions?
This is the reason: https://t.co/VxxNWfhj1o
LLM attention spans are the hidden bottleneck in your coding workflow and will remain so for the foreseeable future. Navigate these constraints and vibe on!

[fnerdman @fnerdman]

Let's be honest to ourselves: is the current AI hype not just a massive timesink?

[0xDesigner @0xDesigner]

my honest review of openclaw as a non-technical:

it requires lots technicals to get set up. api set up, doing things in your terminal, whatever.

but it's the most non-technical friendly assistant i've ever used.

no prompt tricks, no file management, no architecture or database designing thingy whatever.

you can just ask it to change it's programming and it will.

especially with it's "memory" and whatever this new QMD database update is (no clue but i installed).

it remembers things. of course, it can be dumb at times.

but you don't need a "planning mode" or any fancy approaches.

it has a genuinely good personality out of the box and just does things.

i really love it.