sophia01

6 ms·

1.
by sophia01 10mo ago
It's Coca Cola vs Pepsi. Yes some might even say Pepsi has been shown to taste better, but people still buy loads of Coke. Of course the tech savvy enterprises will use the best models. But the plumber down the road doesn't care w
2.
by sophia01 10mo ago
The matter of fact is that big companies (think the usual monorepo business going on in FAANG) don't care about the actual code. The code was never the point of the exercise. Eventually you realize this. Code is like the ether. The com
3.
by sophia01 1y ago
Thats not true. They price the stocks at current market value and tell you how many you'll get + what the vesting schedule is.
4.
by sophia01 1y ago
> The back-loaded vesting schedule is such blatantly cynical bullshit. I don't understand this. A friend was recently offered an insane pay package from Amazon (compared to another big-tech). The way I saw it, the Amazon pay package
5.
by sophia01 1y ago
They're not open sourcing it though?
6.
by sophia01 1y ago
Driver license and selfies. Also still not available in API after doing that! Edit: I do have access now via API.
7.
by sophia01 1y ago
API usage requires organization verification with your ID :(.
8.
by sophia01 1y ago
> The UK is not that different by comparison. Do note that the UK is 15.6x as dense as Finland, and the climate is quite different: e.g. in Helsinki (southermost city) mean daily temperature is below freezing point 4/12 months of th
9.
by sophia01 1y ago
Yeah just checked apparently it is available as a preview (not on main models/pricing page).
10.
by sophia01 1y ago
My understanding is that the coding agents people use can be modified to plug into any LLM provider's API? The difference here seems to be that Cerebras does not appear to have Qwen3-Coder through their API! So now there is a crazy fas
11.
by sophia01 1y ago
When I read "map", I think "map"...
12.
by sophia01 1y ago
Have been using this all morning for some integral-heavy math for my PhD (trying to bound certain analytically intractable integrals). It's a bit hit-or-miss. It's been able to come up with some pretty impressive bounds but also f
13.
by sophia01 1y ago
But the data is... wrong? Google Gemini 2.5 Flash-Lite costs $0.10/mtok input [1] but is shown here as $0.40/mtok? [1] https://ai.google.dev/gemini-api/docs/pricing#gemini-2.5-fla...
14.
by sophia01 1y ago
If this is actually competitive with Gemini 2.5 Pro that would be insane esp for an Apache2 truly open weights model, let's hope it's not too hacked to shine on benchmarks!
15.
by sophia01 1y ago
Oh come on. You are going to debate whether a 3% parasitic tax is good or bad. Interchange is paid on debit card transactions too, it's not credit card specific.
16.
by sophia01 1y ago
The title is disappointing: it implies a causal relationship. The study was an observational study. Indeed, you can outrun a bad diet: we all know that. The study just shows that the lack of activity isn't the main reason for obesity
17.
by sophia01 1y ago
Thank you! I am wondering how narrow the focus has to be for an EB-1A. I have been involved in a wide variety of projects, some receiving a good amount of media publicity (100+ mentions across various media), others adding up to a good numb
18.
by sophia01 1y ago
Thank you Peter, this is super helpful!! Does your firm do immigration law consultations or do you have any that you would recommend to ask some more questions?
19.
by sophia01 1y ago
I'm graduating from my PhD and starting a FAANG research role. I'm Australian and about to start in F-1 OPT status. I am interested in an EB-1 or EB-2 (possibly NIW) visa, but am concerned about a long period where I can't tr