Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
textlapse
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
61.
▲
by
textlapse
2y ago
Basic probability/stats and some basic fundamentals around dynamic programming/recursion would be very helpful. The big problem I found with this field is that the core ideas are very subtly built on top of each other. Without a p
62.
▲
by
textlapse
2y ago
For a layperson, this is an accessible and directionally correct definition. For the HN audience, of course this is 'technically incorrect'. The article was written for the (larger) general public. I am also glad they didn't
63.
▲
by
textlapse
2y ago
I am not worried about the attack vector mentioned in the article - but that’s definitely a concern and a nice buzzworthy headline. I am more worried about the robovac companies really being an ad supported spying company with robovac as a
64.
▲
by
textlapse
2y ago
All of the HW components are commoditized and horizontally integrated anyways. It’s easy to verify and build components with strong guarantees that ensures that the integrator can’t workaround (and why would they?) I do worry that companies
65.
▲
by
textlapse
2y ago
This will also end up breaking up the Ads to Nobel Prize pipeline. I do wish they had some sort of GSoC style programs to do what Alpha Go Fold Zero did for the world.
66.
▲
by
textlapse
2y ago
I could imagine an IR or LiDAR like image that is still useful to detect objects or humans but is otherwise indecipherable by a human… Really I think I am imagining homomorphic encryption style techniques.
67.
▲
by
textlapse
2y ago
What I am talking about is a bit different: imagine if the CCD produced a non MxN color image. Maybe think of it as scrambled data that has just the right level of detail for the machine to do its thing but not something where you can get b
68.
▲
by
textlapse
2y ago
Why is it that a smart device (robovacuum or proximity sensor etc) etc require the same technology as a streaming webcam? In other words, are there any HW-level privacy-preserving CCDs (for lack of a better word) that provide an image in a
69.
▲
by
textlapse
2y ago
Kotlin has the problem of "I had N problems, so I think I can create 1 new solution, now I have N+1 problems" (obligatory https://xkcd.com/927/ ). The 'money/percent/usd' are almost like ne
70.
▲
by
textlapse
2y ago
A cigarette company’s wet dream made real by social media companies. And no physical products (no more marginal profits) nor pesky FDA approvals to go through either. What a time to be alive! /s
71.
▲
by
textlapse
2y ago
Is a control group even possible in this day and age?
72.
▲
by
textlapse
2y ago
I expected the last line of the blog post to have said something like "this blog post was created using the help of Canvas" - a missed opportunity or the product not being there yet? Also is this in response to the recent notebook
73.
▲
by
textlapse
2y ago
This is great. I do worry that a future more sinister malicious patent troll could read all the wonderful strategy Cloudflare used and work around them. Hopefully Cloudflare legal team got stronger! Kudos to the likes of Cloudflare and (yes
74.
▲
by
textlapse
2y ago
Maybe OpenAI is trying to enter a new enterprise phase past its startup era? They have hired CTO like figures from ex MSFt and so on … which would mean a natural exit for the startup era folks that we have seen recently? Every company wants
75.
▲
by
textlapse
2y ago
Using an intelligent algorithm to guide a dumb non-intelligent next word predictor is still a non-intelligent algorithm at the end of the day. Sure it’s sorting through garbage more elegantly but it’s still garbage at the end of the day. I
76.
▲
by
textlapse
2y ago
I hope they modernize and come to terms with the reality of the 21st century workforce… definitely needs a fed level revamp on the former part. The clearance part sounds like a good thing though…
77.
▲
by
textlapse
2y ago
I wonder if this helps NSA in their recruiting process - there might be a bunch of disinterested, bored FAANG folks who could 10x their impact by serving their country instead. Adding to the GDP is probably better in the long run but at thi
78.
▲
by
textlapse
2y ago
A $600M Anti-Chesterton Fence that nobody knows why it’s there and everyone’s scared to touch.
79.
▲
by
textlapse
2y ago
How cool would it be to have a reinforcement learning playground to teach ‘crows’ and ‘octopuses’ and ‘cuttlefish’? The key thing is the biological evolution that seeks rewards for survival and reproductive partner selection over generation
80.
▲
by
textlapse
2y ago
I have been reading the Reinforcement Learning textbook for the past three years religiously- didn’t get the concepts clear until this week, made it past Chapter 3 finally. The excellent YT lectures by David Silver and Emma Brunskill have b
81.
▲
by
textlapse
2y ago
I can almost squint at this and see 'self driving cars over a long enough period of time in reality are just .... trains with cars connected by wifi instead of physical beams'.
82.
▲
by
textlapse
2y ago
Without profiling, such statements carry little weight. In particular, for optimal performance unless you use data oriented design and structure using SoA vs AoS. However for a 2d top-down game where you have extremely good culling and laye
83.
▲
by
textlapse
2y ago
Maybe they went through the conundrum of ‘beautiful, shiny metal boxes - one size fits all’ that sits hidden in some server warehouse or dusty room? I can’t think of any successful, non-consumer facing product that Apple has ever built - it
84.
▲
by
textlapse
2y ago
You mean like UPI? For a nascent industry such as India I imagine you could: but for a legacy infrastructure such as that of the US - very advanced a few decades ago but now showing its age - the momentum is too hard to deprecate and start
85.
▲
by
textlapse
2y ago
All squares are rectangles I guess.
86.
▲
by
textlapse
2y ago
Heard from someone: "C++ is a hammer, but then everything starts to look like your finger"
87.
▲
by
textlapse
2y ago
Context is everything. They had every chance to own up from the day of until now. A ‘lulz haha we goofed up’ in a nerdy security conference doesn’t seem like the right place or time.
88.
▲
by
textlapse
2y ago
This may end up in one of those court evidence videos or lawsuits - this isn’t a funny thing. This would have been a closed moment (just a bunch of security nerds discussing something) but instead this is now freely available for the wider
89.
▲
by
textlapse
2y ago
I wonder if this is how astrology was born. You can draw arbitrary connections between things if you stare at them long enough.
90.
▲
by
textlapse
2y ago
That’s a cool paper - but it seems like it produces better debaters but not better content? To truly use RL’s strengths, it would be a battle of content (model or world representation) not mere token level battles. I am not sure how that wo
More ›