zamalek

7 ms·

121.
by zamalek 6mo ago
FWIW, if you want frontier-level performance, as it was a few months back, Deepseek v4 and K2.6 are there. Almost zero chance you can run them locally, but you do have choice in terms of providers. Qwen-coder-next is considered SOTA for thi
122.
by zamalek 6mo ago
Associating dark matter with epicycles is unfair, but it's still a risk (CDM, WDM, SIDM, and probably more by next year). In that light, of course it has more observations (by virtue of creating branches to explain incompatible observa
123.
by zamalek 6mo ago
Ah good point, complete brain fart on my part.
124.
by zamalek 6mo ago
You only allocate on box futures, which are much more rare than naked futures - generally only used where object safety (essentially dyn support) is required. Even then some workarounds exist. Edit: and tasks.
125.
by zamalek 6mo ago
Tokio is a general purpose async runtime. Much the same could probably be said for async-std (except IIRC they do have a barebones reactor for you to build your own on). In general, a general-purpose async runtime will do worse for highly s
126.
by zamalek 6mo ago
> possibility into a binding obligation I never said "binding obligation," that is the first time "binding" has appeared in this discussion and was introduced by you. Once again claiming things I have never said. Do
127.
by zamalek 6mo ago
> You are strongly implying that keeping the vulnerability secret is following of what you quoted. Please don't put words in my mouth when I have clearly stated the contrary. I used the word "disclosure," that is very diff
128.
by zamalek 6mo ago
That's still extremely different to this in one of the GP comments: > There is no such thing as "the responsible disclosure protocol". And yes, I admit I got dragged down to their level and beat myself with a dumb stick in
129.
by zamalek 6mo ago
> Is that a rule? No, it's commonly followed practice: https://en.wikipedia.org/wiki/Coordinated_vulnerability_disc... I'm all for lighting a fire under the developer's ass, but we live in an imperfec
130.
by zamalek 6mo ago
Grok was supposed to be the uncensored frontier model. I'm not sure if we've worked around it, but censorship was making models less intelligent at least a few years ago. xAI have been caught making it agree with everything Elon s
131.
by zamalek 6mo ago
From elsewhere.[1] > You: No, I wouldn't, because my own preferences are towards immediate disclosure. And there it is. You could have said "I don't think responsible disclosure is a good idea" and moved on, but now w
132.
by zamalek 6mo ago
Tons of distros were not informed.
133.
by zamalek 6mo ago
https://en.wikipedia.org/wiki/Coordinated_vulnerability_disc... > In computer security, coordinated vulnerability disclosure (CVD, sometimes known as responsible disclosure) I guess you can learn something new after
134.
by zamalek 6mo ago
Yes, and that's why we have the responsible disclosure protocol. It wasn't correctly followed here.
135.
by zamalek 6mo ago
The disclosure was more about marketing than security. From the disclosure page: > Is your software AI-era safe? > Copy Fail was surfaced by Xint Code about an hour of scan time against the Linux crypto/ subsystem. [...] > [Tr
136.
by zamalek 6mo ago
Platform support for AI has as much place in a browser as it does in Notepad. This isn't about being open-minded at all. I have written multiple MCPs, I use it daily, I am not in the crowd who "don't have an open mind.&quot
137.
by zamalek 6mo ago
AI massively empowers people who are incapable of anything except bikeshedding. It itself is very likely to be a bikeshed (but there are legitimate uses), and it also gives them to power to drone on until they overpower any opposition to th
138.
by zamalek 6mo ago
Only the weights and the RNG used to select tokens can answer that. You will understand much if you read up on the quality of code in the CC source leak, it's completely vibe coded and the printf fn is genuinely impossible for a human
139.
by zamalek 6mo ago
I love it too, but instantly knew it would be polarizing the first time I used it :)
140.
by zamalek 6mo ago
Ligatures are a renderer issue, so using alacritty as a lib wouldn't have this issue (it does demonstrate their hardline stance). Another example that would translate is how long it took them to support disambiguation of key combinatio
141.
by zamalek 6mo ago
Exactly, only humans should have at least one chance to grow and improve. Orgs are heartless legal entities that deserve no loyalty whatsoever, they are all one acquisition away from turning on you (as a customer or an employee).
142.
by zamalek 6mo ago
Large orgs aim to produce some type of output. Their entire existence stems from a "perverse incentive."[1] Governments produce bills and laws, corps produce short-term profits, etc. I am pretty sure that preventing this type of w
143.
by zamalek 6mo ago
That's the one!
144.
by zamalek 6mo ago
There was somewhat recently a post here about how priorities, pressure, and management subverted Dave Cutler's vision for Azure (which was to have near zero human involvement) - my Google fu isn't strong enough to find it. Suppose
145.
by zamalek 6mo ago
> Our priorities are clear: availability first, then capacity, then new features. No mention of Copilot/slopiffication. Probably an intentional omission as Microsoft only has one true priority across all of its products.
146.
by zamalek 6mo ago
The FX-888D is known to have a significant time delay between tip temperature and reported temperature, which is a newbie landmine. https://www.reddit.com/r/soldering/comments/1le2y4l/comment/... h
147.
by zamalek 6mo ago
Breadboards[1] with breakout boards[2] (the Pi Pico in particular is incredible, having an unusually detailed and correct datasheet). MicroPython is probably a great way for gets to get started on the compute side of things. Once you have s
148.
by zamalek 6mo ago
Awesome! It's been a while, but my next level of learning was designing PCBs without breakout boards (and I had several failed revisions). This will be great to learn from.
149.
by zamalek 6mo ago
Put infra deletion locks on your prod DBs right now, irrespective of whether you use agents. This was a well established practice before agents because humans can also make mistakes (but obviously not as frequently as we're seeing with
150.
by zamalek 6mo ago
Only when required, though. It's a conflict resource (which tends to make it more expensive). Same would apply here, but the paper talks about moving molecules around so maybe quantity would eliminate the issue.

More ›