arnaudsm

25 ms·

331.
by arnaudsm 4y ago
You summarized modern AI : good for cherry-picked demos, not reliable enough for the real world. We need more fondamental research to break that barrier.
332.
by arnaudsm 4y ago
Low-end indeed, I used the AF-P DX NIKKOR 18-55mm f/3.5-5.6G VR. Better lenses are more expensive than my phone though.
333.
by arnaudsm 4y ago
Yes this is a 100% crop to highlight the sharpness is handheld low-light, the worst scenario there is
334.
by arnaudsm 4y ago
When my mother uses both in auto mode, the pictures turn out the same quality. This article is about the general public, not us, the HN crowd which love to push hardware to the limit. Which is the historical definition of hacking btw :)
335.
by arnaudsm 4y ago
I'd love to see a graph of sharpness/$ for both categories ! I think they're equal until the $400 range, after Mirrorless obviously wins
336.
by arnaudsm 4y ago
What's wrong with my age? I've done 6 years of professional photography to pay for college. Portrait shoots, weddings, even produced videos, ads, festivals, wildlife documentaries. I worked on Nikon D4S fullframes. How is my age
337.
by arnaudsm 4y ago
Sure, prosumers like us can squeeze extra juice out of the DSLR. But our mothers cannot.
338.
by arnaudsm 4y ago
They have the same used price, which is the point of my comparison. Yes there's HDR bracketing, but we only care about the result. One is blurry (optics), the other has artifacts (AI), but overall sharpness is similar.
339.
by arnaudsm 4y ago
True dynamic range yes, but smartphones have better HDR bracketing software. So my phone also beat my middle-end DSLR in backlit scenarios.
340.
by arnaudsm 4y ago
The point of my comparison was price. You can find better smartphones too. The A72 and D5500 have the same used price.
341.
by arnaudsm 4y ago
Sure, here it is ! https://imgur.com/a/VyHLYqP Protocol: handheld at 10 PM, 10 shots each, at different ISOs, picked the best one The bottlenecks are different, but the sharpness is comparable. - The DSLR was limit
342.
by arnaudsm 4y ago
Disclaimer, I'm a retired pro photographer that sold his full-frame to focus on software engineering. - The X-Pro 3 is $2k, not what I call middle-end. - I agree on the aggressive AI processing. Fortunately I could disable it.
343.
by arnaudsm 4y ago
Ergonomics is a mixed bag. DSLRs win at latency and burst, as well as manual mode. But sharing the pictures is a pain, the UI is hard for beginners. And the most important ergonomic of all : it's easier to grab my phone than the 1-poun
344.
by arnaudsm 4y ago
I traveled to Japan recently with an middle-end smartphone (Samsung A72 with 12+35+60mm) & a middle-end DSLR (APS-C with 18-55mm), that I bought the same price. Surprisingly, picture quality was on par. Low-light, stabilization, everyth
345.
by arnaudsm 4y ago
I hope this FAANG downturn will push software engineers to new industries, and bring some cross-pollination. What happens when the world's most brillant minds do something else than making us click on more ads ?
346.
by arnaudsm 4y ago
I built an ultralight markdown CMS with this philosophy in mind : https://github.com/arnaudsm/raito Just drop index.html on a static server, and it's live.
347.
by arnaudsm 4y ago
Not everyone uses Twitter for political flamewars. I use it to follow my favorite researchers and innovators. It's the only website where I can actually interact with Yann Le Cun just like I was at a conference. (with HN of course)
348.
by arnaudsm 4y ago
Yes there's plenty of reasons, channel saturation, low quality routers, various interferences. My Ethernet cable is easier to debug.
349.
by arnaudsm 4y ago
When you're on speaker, most platforms have a noise cancelling algorithm to prevent Larsen loops between mic and speaker. Using headphones usually disables it.
350.
by arnaudsm 4y ago
Because controlling the stack allows you to monetize it later.
351.
by arnaudsm 4y ago
My golden rules to make videoconferencing more reliable : - Use ethernet. Wi-Fi isn't reliable, your neighbor's microwave can ruin everything. - Use wired headphones. 0 latency. 0 connection time. And disables the mic noise-
352.
by arnaudsm 4y ago
Indeed, we'd need an interventional study. Forcing a random study group to drink for decades seems unethical and expensive, but I wish we had the results.
353.
by arnaudsm 4y ago
I'm referring to the UI when you're specifically in watch later mode. Imagine if opening a Google result took 2 clicks, it'd be infuriating.
354.
by arnaudsm 4y ago
The Watch later UX is atrocious too : - Adding : hover the video 1000ms to reveal the button - Removing : 2 clicks away under a menu Why hide the 2 most frequent actions? Do you even do user testing?
355.
by arnaudsm 4y ago
It's not only 2D 360, there's full 3D photogrammetry. Click on the dollhouse button on the bottom left.
356.
by arnaudsm 4y ago
What's the difference between chrome's and Firefox's incognito?
357.
by arnaudsm 4y ago
Indeed photos are the gateway drug to data hoarding. I'm at 128 TB myself, I now have Wikipedia dumps, millions of books, thousands of YouTube videos, papers, everything I can download pretty much. You'd be surprised how ephemeral
358.
by arnaudsm 4y ago
We underestimate how latency matters for user experience. Like the famous "100ms = 1% of sales" at Amazon https://news.ycombinator.com/item?id=273900
359.
by arnaudsm 4y ago
Brave is way worse. They have product ads everywhere, pushed a weird crypto scam, and even injected affiliate codes in URLs. Firefox has ads on their new tab too. We need better and more respectful competitors.
360.
by arnaudsm 4y ago
I don't buy the SP500 graph argument, economies tend to behave on a log scale.

More ›