minraws

4 ms·

1.
by minraws 9d ago
Normal People get to think they are helping 3$'s worth, and Rich people get to think they are helping 3$'s worth... Both side get to feel like they got more out of the initial investment. 1:3 2:3 Everyone wins inside their heads a
2.
by minraws 9d ago
I don't see a reason why it should AMD doesn't care about lower end prosumers atm. They might in the future but future is in the future ofc Edit: to be clear I think it's ridiculous they don't but from a company's s
3.
by minraws 14d ago
Unless you throw the investors and senior executives in jail how do you intend to do anything more than performative dance with these measures. I personally think countries should start sanctioning and international warrants for all large i
4.
by minraws 14d ago
They likely were but Companies don't go to jail they pay fines
5.
by minraws 15d ago
1. Reflog is very ref specific that is it tracks head. JJ has op and evo logs that do that + a bunch more they track everything about the repo every possible action. It's hard to explain with going into technical depth but you can read
6.
by minraws 15d ago
jj is one of the few new devtools I was quite frustrated with while starting out but could see the value so stuck around and man is it so nice. For anyone curious the big thing with jj is you can undo. Basically if you were running a rebase
7.
by minraws 20d ago
> how much demand is still gated behind cost constraints. The market for this is HUGE. I think this misses the actual limits here. The problem isn't demand it's, "how much people are willing to spend on it". Cheap AI
8.
by minraws 20d ago
I honestly feel like we centralized too much on huggingface same as github.
9.
by minraws 21d ago
> 2) Hardware costs have continued ascending with no sign of letting off, so it's unlikely that a DGX Spark depreciates to zero in one year. If someone told me that costs for X will keep increasing because they have been increasing
10.
by minraws 21d ago
At 50tps for single stream you are going to get 50 * 60 * 60 * 24 * 30 = 130M out tokens of GLM 5.3 Flash... That's less than what 40$ at current API rates... So if you are willing to pay 200$ per month you will get much better limits
11.
by minraws 21d ago
You aren't going to get nearly as much token usage locally from DGX Sparks or even M5 Ultra (though it might be close, unsure would need to get my mittens on it to clarify). You will get around 2-4 concurrent streams of aggregate token
12.
by minraws 27d ago
Best deal of the century multi-trillion dollar future guaranteed.
13.
by minraws 28d ago
GLM 5.3 (maybe because not open weights) Otherwise I think they just meant chinese models not mainstream models. Like Hy3, Mimo, etc. Also gemini models but not sure if they are mainstream anymore.
14.
by minraws 28d ago
You are in office and then you hunt for a quite office room for every call, within the same campus you hope everyone gets to the same conference room which in bigger companies is infeasible so you are still remote aka face to face over a ca
15.
by minraws 29d ago
Microsoft bet everything on possible AGI, no one thought the result would be a bonafide commit printing machine.
16.
by minraws 29d ago
I feel like I and HN live in different worlds, https://www.blocksandfiles.com/ai-ml/2026/01/20/samsung-sk-h... Big manufactures have absolutely cut production of lower margin products and it's not l
17.
by minraws 29d ago
In an absolutely free market? sure. But we don't live in an absolutely free market, the memory manufacturers have refused to budge on creating new fabs until recently out of caution and to be able to create a demand excess. They have s
18.
by minraws 29d ago
Just to be clear the big Memory giants are famous for their price fixing historically speaking.
19.
by minraws 1mo ago
Because it's convenient? Shader and vulkan semantics can be quite limiting and annoying to write. Maybe it doesn't matter in a post AI world but perhaps it will allow better abstractions. No need to yuck someone else's yum.
20.
by minraws 1mo ago
Pointers are sort of needed for high performance memory management for HPC targets for existing design patterns, maybe we can think of better solutions down the line but it's hard for me to say anything I just use/abuse CUDA point
21.
by minraws 1mo ago
For community projects it's hard to manage user accounts and stuff.
22.
by minraws 1mo ago
I can bet AWS won't be doing any serious foundational model development, I knew folks on their AGI team now it's not even big enough to compete with the chinese labs... :3 Azure I am not sure how long they will be around doing MAI
23.
by minraws 1mo ago
I both think OpenRouter is not worth 7B$ and Stripe might have made a brilliant move in making this purchase, honestly just troubling thoughts around model inference and model token as the new traded currency go through my head. Imagine if
24.
by minraws 1mo ago
There is more than 1 fatal airplane crash per month in UK? Where are your stats for this? Heck there isn't more than 1 commercial airplane crash in every few years globally. Trains might be higher but if UK is having 12+ crashes a mont
25.
by minraws 1mo ago
Is there an opt-out mechanism at least? CF is burning goodwill in months it built over the last decade.
26.
by minraws 1mo ago
If trains were having fatal crashes every month I will and I think I speak for everyone of my friends in UK, they should stop using the train. Same with airplane, but often thats the only solution if cloudflare is your only solution in your
27.
by minraws 1mo ago
This might be rude but this is the reason we have shitty software everywhere. The "there can always be a bug, get over it mindset" is the reason software sucks, everything needs a dozen patches to even work and no one can have any
28.
by minraws 1mo ago
That depends on when you are talking about... Compared to modern social media platforms they are eons behind ofc social media has more content and more data ofc. But their BigData and stuff was bleeding edge in pre-2018 in 2026 they really
29.
by minraws 1mo ago
Bugs happen all the time yes, outages don't. If outages are increasing with scale you get 1 or maybe 2 free passes. After that you either have in-ept Engineering or just in-ept leadership. I used to be all in on CF a while ago, now I a
30.
by minraws 1mo ago
Now they do have SpaceX's GPUs and Grok models... I mean I get that we can all dislike this move but it might the best way to get into the race to become one of the top labs. Ofc Grok 4.6 is not there yet, I have seen people claim it&#

More ›