throwa356262

10 ms·

1.
by throwa356262 5d ago
I think the bigger problem is that this was pushed as a firmware update, bricking previously working devices. Remember when FTDI did this and suddenly medical equipment at some hospital stopped working?
2.
by throwa356262 5d ago
I think some people are missing the joke here https://en.wikipedia.org/wiki/Npm_left-pad_incident
3.
by throwa356262 7d ago
Dont know about that. A significant portion of the posts are of the type "new version of model X was just released" or "I just built something by prompting an LLM a few times". Latest papers from DeepSeek and DeepMind on
4.
by throwa356262 8d ago
"I also now work at Canonical..." You know the guy doesn't easily give up if he has gone through Canonicals famous 10-20 round interviews.
5.
by throwa356262 8d ago
Hold on, what happened with grok?
6.
by throwa356262 8d ago
To be fair, the GPU in pixel 9 and 10 is pretty bad. I really hope they stop using PowerVR in pixel 11.
7.
by throwa356262 9d ago
Yes, but SF salaries are still twice German salaries and probaby 4x Eastern Europe salaries.
8.
by throwa356262 9d ago
Any idea how much this has cost them? Some non-profits have their HQ in the valley and pay SF salaries when they could just as well have paid a fraction of that to a very competent developer elsewhere.
9.
by throwa356262 9d ago
"27 to 30 December 2026" I have always wanted to attend CCC in person, but these dates for the main conference only work if you are 20 and single...
10.
by throwa356262 9d ago
"We implemented a series of aggressive memory optimizations, including..." This whole thing sounds like industrial scale auto-research, but done by people who actually know what they are doing.
11.
by throwa356262 12d ago
Proposal: Set up a mastodon account that mirrors your (or someone elses) X account. This way people who want can continue using X, and other people can read & refer to their posts without needing to log into X.
12.
by throwa356262 12d ago
Importing is usually not an issue, as you can always enter the secret manually. It is the exporting that is the problem. The secret looks something like this: JBSW Y3DPF QQHO .... (usually fairly short unless its google)
13.
by throwa356262 12d ago
The issue is that Android is open while Apple seldom talks about their security issues. This might make Apple look like the more secure option but the reality may be different because Android is more scrutinised. If you are sceptical cons
14.
by throwa356262 12d ago
Some apps such as Aegis allow exporting the MFA secrets.
15.
by throwa356262 15d ago
No, because the software layers between GeekBench and the hardware are not identical on all targets.
16.
by throwa356262 15d ago
Reminder that despite their claims, geekbench results cannot be compared across platforms. They don't only cover CPU and GPU but some OS components that can look very different on different platforms.
17.
by throwa356262 15d ago
Opencode system prompt contains a lot of stuff but even worse is oh-my-pi where their long prompt looks like random garbage hallucinated by a 2023 LLM: https://m.youtube.com/watch?v=c_fQoDkULl0 (see around 8:00)
18.
by throwa356262 16d ago
For the first time ever, and that for just a short while. And after significant price hikes that has had their biggeat customers looking for alternatives.
19.
by throwa356262 16d ago
DeepSeek, minimax and so on have razer thin margins but unlike openai and Anthropic they are actually making some profit. Doing this doesn't make any financial sense. Maybe Anthropic is confusing Chinese AI providers with token resel
20.
by throwa356262 16d ago
And Adobe? No hold on... Adobe is probably already secretly owned by BS given their business practices.
21.
by throwa356262 16d ago
Fake it till you make it?
22.
by throwa356262 16d ago
Can we get a normal landing page please? These AI generated pages make my eyes bleed. Anyway, I find the tiny C3 to be much more stable than other versions. It is a bit sad the very few projects support it.
23.
by throwa356262 16d ago
I think they urgently need to free the compute power currently wasted on the big pro models.
24.
by throwa356262 18d ago
On the contrary, small efficient models using recent improvements should be more sustainable than huge frontier models that require billions and years to train
25.
by throwa356262 18d ago
I think you are confusing Apple and Arm. Mali drivers are contributed by Arm: https://github.com/torvalds/linux/blob/master/MAINTAINERS
26.
by throwa356262 18d ago
This Xiaomis second or third chip ever so I assume there will be problems.
27.
by throwa356262 18d ago
My reasoning is that we have seen multiple small labs come out of nowhere and create impressive models.
28.
by throwa356262 18d ago
For the first two, I highly suggest you watch a few of their videos from about a year ago.
29.
by throwa356262 18d ago
I don't know. The moat doesn't seem to exist. Probably safer to stay out of the race for now.
30.
by throwa356262 18d ago
Why? I thought these GPU were open source with mainline kernel support?

More ›