HN user

rys

620 karma

https://rys.sommefeldt.com/

Posts5
Comments195
View on HN

He does explain it in his blog post. He changed it after the erratic communication and actions of RC leadership, then after realising what they were really doing, left them to complete their “security audit”, assuming they’d discover it themselves and take appropriate action as part of that. That never happened (which is wild), so he let them know.

They still don’t seem to be in complete control or understanding of the infrastructure they forcefully took control of.

In practice, what gets labelled as the L1 cache in a GPU marketing diagram or 3rd party analysis might well not be that first level of a strict cache hierarchy. That means it’s hard to do any kind of cross-vendor or cross-architecture comparison about what they are or how they work. They’re highly implementation dependent.

In the GPUs I work on, there’s not really a blurred line between the actual L1 and the register file. There’s not even just one register file. Sometimes you also get an L3!

These kinds of implementation specific details are where GPUs find a lot of their PPA today, but they’re (arguably sadly) usually quite opaque to the programmer or enthusiastic architecture analyst.

One interesting thing is they seem to have also cloned the number of stars each repo has on GitHub. At least that’s true for a cursory check of a few that I’ve committed to. If you click on the star count you’ll then see two separate counts for GitHub and GitCode stars.

It’s one thing to mirror the repos, but it’s another to initially misrepresent the interactivity the repo is getting, even if they’re clear about it when you dig in.

Also, how do you authenticate so you can keep committing and interact with the repo to manage PRs and issues if it’s yours?

So was there ever a deal with OpenAI? Nothing in the keynote mentioned them or needs them. If there isn’t a deal, I’d love to know how everyone claiming it was signed on the dotted line was led so far down that garden path.

UniFi U7 Pro 3 years ago

Just in case anyone reads the parent and the takeaway is you can’t: you can still operate UniFi that way if you want to. The cloud connection and apps are optional.

Agreed. As to the world lived in: most people are so enthralled with money that any risk to earning more in the future is unacceptable to them, sadly. Companies then bank on that behaviour to help sweep things under the rug.

The mention of OLED is sideways to the network stuff, it just so happens the device he chose has an OLED screen.

The protocol is FreakWAN, so there’s no TCP/IP involved. LoRa devices send messages opportunistically and listen out for other messages they can understand and then maybe act on if they choose to. It’s about as simple as you can imagine.

One thing that struck me as I read through was that the computational cost for TLS 1.3 is so expensive that it becomes infeasible to run on the slower machines, resulting in timeouts and failed connections while setting them up.

Yet those old machines had full GUI OSes and performant apps and snappy interactivity in their day. Is the sheer amount of computation required to facilitate TLS 1.3 really more than what was needed to run entire interactive GUI operating systems of that era?

Kind of mind boggling if so.

Maybe it wasn’t meant to be taken literally, but it’s difficult to believe he listened to tens of thousands of tracks, which would amount to almost 10 weeks of non-stop music on the conservative end of an estimate.

Shorting Bitcoin 5 years ago

I was unsure and looked it up, so in case it helps someone else: the person that died in Costa Rica is Mircea Popescu.

I’m willing to see what Ubiquiti will do to make it right before I switch away, because I have a local-only setup of EdgeRouter and UniFi APs that’s been absolutely great in the years I’ve had it, but this is really last chance saloon stuff now.

I’m looking for a proper post-mortem and the steps to make sure it can’t happen again, recommitment to local-only users and respect of the customer, and a step back from the push to cloud everything.

I bought a domain to use for my home network instead. I have LE issue a cert for it via DNS challenge and use it liberally with hosts on my LAN, with the excellent benefit that I don’t need to give clients a new CA I invented.

Graphcore are coining the initialism, therefore I imagine they’re not keen on it being applied to anything other than their design and its novel architectural approach.

It’ll likely end up just like GPU though, as part of the industry’s nomenclature.

Just have a look at the 10-Ks. Nvidia's: "Gross margin for fiscal year 2019was 61.2%, compared with 59.9% a year earlier, which reflects our continued shift toward higher-value platforms". Datacentre is their biggest growth area today, and they have traditionally created specific high-margin products to extract value in areas with very limited competition.

The same thing is true for Intel when they don't have competition. They are the company they are today due to insane gross margin on Xeon and the explosive growth of cloud computing. DCG has been Intel's top performing BU for ages until recently. Expect the same story from AMD as it eats Intel's lunch in that market. High margin semiconductors are money making machines.

Brit here, and in 2017 I bought a window venting small ish unit made by Amcor. £199 or so, and worth every penny these last few years. When I don’t use it, it lives in my spare room tucked away and I wheel it to my office for the hot spells.