* using living artists work to train models = good
* generating living artists work using said models = bad
Good ethical consistency from the OpenAI crew
3,785 karma · joined August 27, 2014
* using living artists work to train models = good
* generating living artists work using said models = bad
Good ethical consistency from the OpenAI crew
"Pegasus" is not one hacking entity like most articles make it out to be. Its
1) A bunch of services that download data, given root access to a phone
2) a bank of 0-days, we don't know how deep.
For all we know, there are times when "Pegasus" doesn't work for hours, days, weeks, until the 0-day is rotated. We do know from some leaks that they have a mix of non-click and click exploits, and also support all different kinds of phone OS.
Their hacking abilities are definitely overstated, for all we know, for smooth continuous customer support, they could be buying 100% of their 0-days, and not finding any themselves. A 0-click 0-day for iPhones is worth about $2,000,000[1], a company with contracts like NSO can afford a lot of those. IMO the media portraying them as super-hackers is pure hype. Its a bunch of crooked business people who figured out how to extract money out of countries
[1] https://arstechnica.com/information-technology/2019/01/zerod...
I posted the recode article about Amazon admitting to it above
An Amazon spokesperson confirmed the setup and said that Amazon staff is asked to help out when the system used in the new Amazon Go store can’t make a determination
https://www.vox.com/platform/amp/2017/1/6/14189880/amazon-go...
I work in the industry and the 100% success rate touted by these companies just doesn’t exist in live in person real time inferencing. We will get there one day, but in many cases “magical” computer vision AI has humans verifying and changing AI outputs
Imagine Amazon was flexing the stores being fully automated before “Track Everything Everywhere All At Once” was written
Yes, AI powered software can help them associate items to a person, but not even close to the 100% success rate companies like Amazon and Trigo claim
If an autonomous driving car is taken over by a hacker, and starts running people over, how fast would you expect the police to block it/shoot its tires?
"Then collapse" - nope, then generate tens of billions of profit per quarter. Im sure the people of Hugging Face would love that
* Potentially do it better/purer
* Get a few news articles, and a mention in a very high traffic Wikipedia page
* Again if they do it well, get 8 figures in funding pretty easily. While during any other day they might be fighting for $50,000 here and $100,000 there
You could probably buy something that would heat your seat at home for under $4 on Temu. And that includes multiple middle companies and shipping across the ocean. It probably costs them pennies, where the upside is, this owner doesn't want heated seats, but a car can easily have 2-3 owners in its first 10 years. maybe the 2nd and 3rd owners will want the heated seats, worth the money it would take to install it
That's like asking Intel to fix a processor you overclocked
on the other hand its hard to blame one aspect for "being bad" at leetcode. Plenty of non-diagnosed people with tons of real world experience have a hard time finding the gotcha in a 38 minute interview. It doesn't represent the real world in any way IMO
In other words getting an offer for "tech in a non-tech company" is pretty easy. Getting the Meta offer in 1-shot of 6 back to back interviews, not so much
There are plenty of fully remote openings for senior software engineers looking for $100k TC right now. Every person can make their own calculation
Saying all that, if you read 1000 characters, according to the TikTok way of sorting, you're already hooked. There will be plenty of posts where you swipe on after ~10 characters (in terms of swiping on within X hundreds of a second)
People who regularly use TikTok are used to the swipe up super quickly once a post doesnt interest you, it shouldn't be any different for measuring text
So to be more precise, building a product on top of an unprofitable product means things will usually change for the worse (cost cutting) sooner or later
By now this is actually funny to read. Never rely on another companies product to make your own product, without accepting things can change overnight and shut you down
As Llama2 is self hosted, you can choose which iteration to host. Much better developer experience
Edit: to be clear OpenAI is unprofitable, so is Reddit, so was Stadia. Building on top of someone else's unprofitable product is doomed to begin with
This sushi we're having for lunch must also cost a fortune
Jokes aside, this screams of an org that isn't able to judge its employees by what they deliver, so they have to go chase processes
Either Shopify employees are delivering what they're expected to or they're not. It feels like executives aren't able to figure out if they are
* WARNING: 26 weeks parental leave will cost us $xxx,xxx
* WARNING: Buying team lunch this week will cost $x,xxx
* WARNING: Offering laid off workers 20 weeks severance will cost $xx,xxx,xxx
* WARNING: Offering your direct reports raises will cost the company money!
And while we're at it. Should we do performance reviews where a team sees how much income they personally generated vs how much the team costs the company? That would be super fun, that will probably stop most people from asking for a raise
Just ask people to limit meetings to the minimum, no need to flash numbers into a guilt trip for a bank account I do not own. If you asked me my employer should give me $100 for lunch every day, 4 day workweeks at 8 hours a day, 52 weeks parental leave, and other costly things I really don't mind them paying for. The cost of a meeting is the last thing I care about
If I see 2 employees talking about yesterdays football game in the kitchenette, should I report them wasting $1,500 of the companies time to HR?
Oh, and finally, im sure these calculations are based on 40 hour work weeks. Is this a good reason for an employee to tell their boss at Shopify they will not be online or answering any slack/email after 5pm every day?
Everyone should listen to the Reply All episode about the TikTok algorithm, its pretty mind blowing
Back to the main subject. For me the obvious buyers in terms of makes most sense are Meta -> OpenAI -> Shady data broker. Sam Altman is also one of the investors in Reddit
For Meta it goes beyond the data, which is a plus. There are a ton of similarities between FB groups and Reddit. FB groups is even more popular in terms of DAU, and some communities are bigger there than on Reddit. Plus they also utilize free moderation without any uprisings. Owning both communities would be huge for Meta in generating ad $'s
Sure