WinLychee

14 ms·

61.
by WinLychee 3y ago
Upon revisiting this, you make a fair point, I suppose that would drive demand.
62.
by WinLychee 3y ago
Short answer: Yes I feel that definitely, not really for me anymore. We took things way too far and it's pulling us apart at the seams. Longer answer: We're leaving the early(ish) days of the industry. Tech is well established and
63.
by WinLychee 3y ago
At least in part it's not just supply and demand, but what capital owners have deemed your value is worth, and how much value you can produce for them given your role. A programmer at a big co can take that 300-500K/year you pay t
64.
by WinLychee 3y ago
Keep trying, getting probably a 10% response rate over here (not that bad for cold applications), submitted 100+ so far. It can take a few weeks to get back to you. Also, competition is FIERCE especially for remote roles.
65.
by WinLychee 3y ago
The bar has been raised. Not like a few years ago when they were hiring zero experience bootcamp devs (much to respect to these folks who are hard workers from non-traditional backgrounds). Questions are not easy, you can be tossed out for
66.
by WinLychee 3y ago
There's been a large influx of people chasing dollars and status, they could not care less about the product or who uses it. You can still find orgs that do care, but they're swiftly pushed out by the VC-funded "growth at any
67.
by WinLychee 3y ago
Truly beautiful design, but you can also pick up vintage gear for cheap at a thrift store or garage sale, would be close enough.
68.
by WinLychee 3y ago
The same feeling of tech 10-15 years ago is what the current AI scene feels like. There's definitely an emphasis on higher education now, but outside of that look at what hackers have been able to accomplish just solo. It's a burg
69.
by WinLychee 3y ago
Agree that the risk is overblown. Apart from the sometimes frantic pace, startups also have more flexibility for different personalities. In a small company things aren't standardized and there's room for uniqueness. #4 to me is u
70.
by WinLychee 3y ago
At the moment it's slim pickings out there, so even if you have to grind out a few years, it's better than nothing. Some people also enjoy the hectic environment. There're definitely ways to do startups that don't involv
71.
by WinLychee 3y ago
Beans and potatoes are fairly easy and fun to grow? They are very "set it and forget it" and produce a lot in one season. Although if they get infected with some kind of fungus or virus, that could put you at risk of starvation if
72.
by WinLychee 3y ago
In the quest to placate shareholders, one of the failing FAANGs will make a panicked announcement they have partnered with Boston Dynamics, to ensure AI is democratized and equitable for all (technically true when we're all on the run)
73.
by WinLychee 3y ago
There are a lot of financial incentives pushing engineers to seek novelty, even if they might not necessarily want to. Promotions and raises are heavily tied to projects introducing or inventing shiny new tech. Heck, even VC funding for sta
74.
by WinLychee 3y ago
As far as FP langugages go, LISP is very much a free-for-all. It can be pure if you want it to be, or stateful and imperative if you prefer. Nothing is really enforced like for example in Haskell. For your use-case a mutable reference behin
75.
by WinLychee 3y ago
This series of articles from Pinecone seems accessible https://www.pinecone.io/learn/vector-indexes/ . Stepping back a moment, the problem "I want to find the top-k most similar items to a given item from a c
76.
by WinLychee 3y ago
You can take a network and its weights that someone else trained, and use that pretrained network to train on your own data, which is likely to be a better starting point than random weights.
77.
by WinLychee 3y ago
> The parameters θ of the deep neural network in AlphaZero are trained by self-play reinforcement learning, starting from randomly initialised parameters θ. Games are played by selecting moves for both players by MCTS, at ∼ πππt. At the
78.
by WinLychee 3y ago
On Linux this was solved for me by using `tlp` and I check usage once in a while with powertop. It's going to be not so great out of the box (~5-6 hours) but after some basic tweaks I can pull 10. Mostly the fixes are about getting the
79.
by WinLychee 3y ago
Just a thought, but perhaps it's the small community and independent culture of Julia that has led to the high quality of its software. Small groups of highly passionate people can accomplish a lot! If I recall the history correctly, s
80.
by WinLychee 3y ago
Watching that interview, I got the impression that Geoff is a very curious person, driven by his sense of wonder. At the same time I couldn't help but feel that he comes across as very naive or perhaps innocent in his thinking. While h
81.
by WinLychee 3y ago
Thanks for organizing this information into one place! It's super interesting and really relevant to high performance software. A lot of this material is scattered between conference talks, obscure forums, and hacks in closed-source so
82.
by WinLychee 3y ago
One point I would add is that this mostly applies to the "1st world" IMO, because in many parts of the world there is local produce and meat and it's often part of the culture to grow your own (or at least to keep a few chi
83.
by WinLychee 3y ago
Logically that would make a lot of sense, if we were to design things in such a way that products are only shipped to their local urban areas. However, the modern day system has given rise to such absurdities as fruit grown in South America
84.
by WinLychee 3y ago
Depending on YOE you still have a lot of room to grow in terms of compensation (see https://www.levels.fyi/ ). You don't even have to go to FAANG. The hard part IMO is getting an interview, which is solved by getting
85.
by WinLychee 4y ago
Knowledge work in the US (especially tech) is on another level entirely. You would not be treated this way in retail, service jobs, or at most employers AFAIK (this also depends heavily on the state). As far as I know most employment contra
86.
by WinLychee 4y ago
Microsoft used to be able to create really great desktop software. I thought Windows 98, XP, and 7 were incredibly slick and well done. Old versions of Office ran well on CPUs slower than those found in modern toasters. I too am curious of
87.
by WinLychee 4y ago
I'd like to see a real study done on this, but if I were to hazard a guess, I think it boils down to: organization size, feature prioritization, development process, and finally lack of knowledge or care. It's a symptom of the way
88.
by WinLychee 4y ago
All of the listed "facts" or "features" can be incorporated into an index (e.g. with Lucene) or fed into ML-models (logistic regression, trees, neural nets, etc). Without going into specifics, I've seen ranking trea