HN user

chamanbuga

357 karma
Posts0
Comments54
View on HN
No posts found.

This is cap. I worked on heads up glasses, and one of our issues was the lack of integration with Apple's iMessage ecosystem. Device makers are willing to go through several security measures, like deploying the MFi chips and certification. However, at best this gives you access to the notification system, not iMessage itself. You are able to respond to messages via the notification framework, but not integrate directly with iMessage even after taking all security and certification efforts. This isn't a security play. This is a walled garden play.

While the folks you mentioned advanced deep learning techniques, Fei-Fei Li transformed computer vision by creating ImageNet, which fuelled progress in the computer vision field. Beyond that, she’s been a champion of ethical, human-centered AI and has worked to make AI more accessible.

The first guess was shockingly spot on. Tried it again and it thought I was English. Tried it again and it thought I was Dutch. First guess was right though.

I vividly remember this incident and more like it. I was coached by my manager to never reply all to an email addressed to a larger distribution group because of this exact reason. Not only were you annoying people, but everyone's email ended up getting delayed for an hour or two.

Can someone explain to me why the backlog would happen? Why they didn't have systems to protect from such a basic DOS attack?

I disagree with the unbiased part of the comment. IMO all social medias, whether it's Instagram, YouTube, or TikTok - they are all heavily censored and their algorithm is tipped in a direction that optimizes attention, click throughs, and in general, is not good for you.

Notice how on TikTok, even accounts you follow sometimes will show up with the + sign, begging you to click Follow eventhough you are already following them. This isn't a bug, rather, it's dark UX tricks to keep you trained on clicking the follow button.

On Meta (be it Facebook or Instagram), you only need to be sympathetic to what people in Gaza are experiencing to begin to see the censorship. It's blatantly obvious.

Coming back to TikTok, similar to Meta, they have crafted an incredible algorithm that will keep you glued to your phone. We have no idea if they have adjusted the levers to increase dissent in young people in the West. Just because they are saying they haven't, or we have not been able to prove that they haven't, doesn't mean they won't. I can tell you with a guarantee that if Instagram or YouTube were allowed to operate in China, they would certainly use these levers to increase dissent against the CCP. We already have an example of what these companies are doing to censor certain content and boost others in places like India and Pakistan.

In general, all of these social media companies are this generations cigarette companies. We need heavy regulation from congress. We need to take tech's lobbyist money and not do their bidding.

We need to shutdown lunatics like Andreessen Horowitz and more.

This is what I initially thought, however, I have already witnessed working demoes of 3DGS when using a single viewpoint, but armed with additional auxiliary data that is contextual relevant to the subject.

I learned about 3D Gaussian Splatting from the research team at work just 2 weeks ago, and they demoed some incredible use cases. This tech will definitely become mainstream in camera technologies.

IMO Alexa is just a bad product, which Amazon thought scale scale scale instead of finding value and product market fit. So many things that it does is better than predecessors, but overall is subpar and didn't get better over time.

Ex. Music. Bread and butter for Alexa. Conversion opportunity to Amazon Music. Conversion opportunity for podcasts, news, so much... Try asking Alexa to play new music, something different, slow jazz, I didn't like that song, play something like blah, etc... Most of these queries do not work. It's such a POS.

Ex. Proactive notifications. Alexa routinely informs me that diapers or something else from my routine purchases is likely running out and I should refill them. I've used this feature 100s of times in the 7+ years I've owned Alexa. I'm sure the LTV I've given Amazon is hell a lot more than most Alexa users. That being said, I can get so many more recommendations than just the 1-2 things in my routine purchase history.

I could go on and on. Ultimately, like many products by large companies, Amazon had a product that received early success because of novelties like hands-free music, alarms, timers - but never found more utility beyond that. Instead of doubling down on discovery, they worked on scaling out the platform and opening it up to 3P developers. The platform they exposed was limited in frustrating ways, and ultimately did not help them discover any hidden gem skills.

IMO the smart home virtual assistant is ripe for easy disruption. Throw out Amazon's model and use ChatGPT / Anthropic instead. Keep the 3P plugins. This fusion would be a sufficiently better product than what's in the market today.

I have 15 YOE and today I make what I did when I was in Year 2. I have the same employer as Year 2, but I relocated from US, PNW to Canada, Ontario. This resulted in a drastic reduction in pay. After taxes, I make 42% less than what I made before. I make a little more than what I earned as a fresh-out-of-college new hire.

While I'm annoyed about the money, I have an improved quality of life in Toronto. My wife and I were raising kids alone in Seattle whereas we have a village here to co-parent with. I also found much more cultural identity and roots here in Toronto compared to Seattle. I also prefer the climate and politics of Toronto more than PNW. I certainly miss the hikes, but we make that up with non-stop beach trips.

If I can achieve the same financial security here in Toronto that I had in Seattle, I would never move. Unfortunately, this has been difficult to achieve, but I'm confident I can get there.

I used to think the trades is where the money is. I've met so many plumbers, AC repairmen, general contractors, firealarm installers - who are really doing well for themselves. Then just the other day there was a Reddit thread about lack of jobs in trades in Toronto. I did a double take. The joke in our friends was our immigrant parents pushed us to study and get corporate jobs and all the trades folks and real-estate agents seem to be doing better than us. A true grass is greener on the other side moment.

I learned that at most trades companies, there are folks with over 20 years of experience and then newbies. No one in between. Every economic cycle, the folks with 5-10 years of experience get laid off. Apprentice are rarely trained by the old dog and put out in the field with minimal experience.

Perhaps this is a Toronto, Ontario problem. Ontario is going through a major economic cycle that has almost everyone feeling quite hard times. Or perhaps this is just the negativity that is Reddit. There is also the BS that income waged workers pay double, if not more, taxes that businesses or self employed workers.

I see your point. I too have been asked to do something unexpected around end of day that ends up taking the night. Over the years, I have learned to say no. It was difficult to learn, but now people understand how to work with me. You're right, inaccurate to say this is not common, I just don't like how the usecase is promoting the worst stereotypes about managers. There's a lot of positive stuff. Ex. you want down time to spend with your family. There are a few calls you think you can skip on. You can ask Duet AI to attend on your behalf and give you summaries at the end of the day, etc...

AI is here to help us work smarter. Hopefully this creates time to work on more interesting & impactful problems thereby improving productivity. The worst case scenario is if people just end up working more on meaningless tasks.

Remember the "let me google it for you" meme? Not very different if all the manager has to do is ask Duet AI to pull together this report instead of asking you.

I'm digressing. This use case really triggers me.

Nothing irks me more about the productivity space than the first use case on the blog post:

"Imagine you’re a financial analyst and you get an email at 5 PM from your boss asking for a presentation on Q3 performance by 8 AM tomorrow — we’ve all been there"

1. We haven't all been there. This is not a common use case at all. 2. This is toxic AF. We shouldn't normalize this behavior. Someone does this with me, I will politely tell them to F off instead of relying on Duet AI to have it done before dinner. 3. There's no way Duet AI can handle all the heavy lifting for such a major document without human intervention. Whereas the use case of getting a meeting summary does not need as much human intervention. It's interesting they start with this case.

There are a lot of promises here as well that require a lot more infra than just switching on Duet AI: "And because sometimes it’s hard for remote participants to see everyone in the conference room, or their colleagues appear far away and out of focus, we’re rolling out dynamic tiles and face detection that give attendees in a meeting room their own video tile with their name"

You need people to opt-in to facial reco. You need potentially multiple camera angles, or minimally high quality cameras that allow to to perform the crop and optical zoom into people's faces. You need in-room attendees to sign off on the digital zoom into their faces. People don't switch on their cameras at home because of this, yet in the room they will be ok? Not so sure about that.

All in all. Great feature availability. But it needs a balance of realism and how this will all come together.

My take on this was user generated content = content generated by people you follow vs network generated content = content generated by people that you don't follow that the network recommends to you.

There's a huge difference between TikTok's FYP and Following Page and Friends page. While content is all user generated, I expect something different from each of these three modes.

Spoke to a few farmer friends. They tell me that large scale operations are very technically complex and have many digital systems deployed for remote management. That being said, a lot of the numbers here seemed to be miss reported. This seems like a combination of genuine issue and fear pandering.

This is an incredibly click bait article by gifted journalists. Most of these folks are fine and still have similar core priorities. Also there are multiple other responsible AI teams in MS. I have extreme empathy for those impacted, but this is trying to generate news where there isn't.

As of late I keep hearing/reading/seeing that the original author of Terminator and the Matrix is Sophia Stewart. It seems even the FBI ruled in her favor, yet both wikipedia pages have not mentioned her at all. Is this a case of really good fake news, as of late, or Wikipedia simply hasn't caught up?

Snap will continue having an existential crisis as long as their platform is dependent on companies like Apple and Google. They are a camera company, and they need to use their expertise to beat Apple at the camera game, which is no easy task.

Holy, while in Toronto I set the thermostat to 68 and complain if I see it cross into the 70s. Thanks to the wind tunnel I'm in, I'm able to turn off the AC every 2-3 days for 1-2 days. Conventional wisdom tells me this is more costly than just keep the AC running all the time at a stable temperature, but, I don't really care about the cost saving over personal preference.

Title is click bait. Coke Studio has done this several times. Mainstream Pakistani songs are quite different from Indian songs, however, both populations can understand what is being sung. Mixing the two together results in incredible music. Pakistan and India still hate each other. No music is about to unite them after 1000s years of history and polar opposite religions.