mliker

9 ms·

61.
by mliker 4y ago
if you enter an entire patient history, it could easily be an identifier of the person whereas Google queries have a smaller max limit number of tokens
62.
by mliker 4y ago
Very interesting read. The author’s point about traditional academic conferences not accepting papers for GPT3 seems plausible given that Meta AI Chief Scientist Yann LeCun tweeted about why ChatGPT wasn’t impressive
63.
by mliker 4y ago
Yes, but the point is the headline is clickbaity, and for the uninformed masses, it sounds like it's highly likely it's from the lab, when that judgment was made with "low confidence"
64.
by mliker 4y ago
Went through the syllabus. The content is really useful since I spent most of my post-graduate years reading man pages for various debugging tools.
65.
by mliker 4y ago
It's interesting to note Marc talks about the future sizes of hypermedia in this other 1993 thread on the HTTP2 spec: http://1997.webhistory.org/www.lists/www-talk.1993q1/0002.ht...
66.
by mliker 4y ago
Dr. Sara Gottfried talks about a slate of tests in her recent podcast episode with Andrew Huberman: https://hubermanlab.com/dr-sara-gottfried-how-to-optimize-fe... While the podcast is focused on female health, the advice a
67.
by mliker 4y ago
this is a pretty biased and uninformed opinion. Pretty condescending to call it a "bunch of dorks...running around unsupervised."
68.
by mliker 4y ago
> doesn't democratize LLMs open source always leads to more democratization in the long run
69.
by mliker 4y ago
People’s contempt rose because the media painted these platforms to be “evil”.
70.
by mliker 4y ago
there is such a thing as measuring too much, but looking at just "revenue" and watching it go up without understanding why and without looking at peripheral metrics isn't true understanding and leads to suboptimal results. Al
71.
by mliker 4y ago
how in the world do you draw such a conclusion from an otherwise interesting and advancing piece of technology. every tech has drawbacks, we can't hold back as a species just because there are drawbacks. we need to build around them.
72.
by mliker 4y ago
it was the most successful marketing campaign targeting developers. And now you're writing about it. I would say that's anecdotal evidence of success.
73.
by mliker 4y ago
We have geniuses like the inventor of CRISPR, penicillin, deep learning, etc.
74.
by mliker 4y ago
You should inform yourself of basic knowledge.
75.
by mliker 4y ago
In absolute terms, there's no rubbish, but yes, it hasn't peaked. In relative terms, the rubbish is about the same. There are simply way more videos than ever before.
76.
by mliker 4y ago
2012 is around when they rolled out personalization of your feed, which drastically increased hours spent on YouTube. So the data shows it was on an upward trajectory since then.
77.
by mliker 4y ago
Keep in mind Meta isn't developing just 1 product. The metaverse needs advanced tech for it to be realized. Once one or two companies develop the tech, of course new competitors will come in and possibly do it better. But you still nee
78.
by mliker 6y ago
> I am (very slightly) more surprised that they still think Facebook advertising is a good (in the sense of "effective") use of their money, given all that has come out in the last few months about how poorly targeted FB's
79.
by mliker 6y ago
These companies did "release open data". It's called Cambridge Analytica. Look it up.
80.
by mliker 6y ago
No, that's not the way it works at big companies like AWS, Google, Microsoft (source: me, I've worked at all 3). A large company attracts various types of people. It's not homogeneous like what you describe. Also, there are p
81.
by mliker 6y ago
Airflow can be used to process data that only adds up to 10s or 100s of MBs. In aggregate, a big tech company processes petabytes or exabytes, but remember, technology like Airflow isn't processing all of that simultaneously within one
82.
by mliker 7y ago
It doesn’t “sound like” that, it’s literally what they’re admitting to. Let’s not spin the narrative here on HN that Zoom didn’t admit to their faults.
83.
by mliker 7y ago
This "chat app" is used by more than a billion people in the world.
84.
by mliker 7y ago
That law would need to be applicable to all companies - Walmart, Target, Visa, etc.
85.
by mliker 7y ago
It's very much context-based. The cofounder doesn't have to give up shares of the company if it's a new company.
86.
by mliker 7y ago
No, that's not the parent poster's point. From parent post: > If they had just let disabled individuals get cyber bullied they could have avoided this negative press altogether.
87.
by mliker 7y ago
Yes, but then the disabled individuals would have been mentally traumatized. Companies shouldn't avoid doing things just to avoid negative press. They're providing a service for the people.
88.
by mliker 7y ago
> that is a power they don't want to lose What makes you think they want that power in the first place?
89.
by mliker 7y ago
Agreed. Facebook came out with the Newsfeed, and now nearly every social product has a newsfeed: MyFitnessPal, Twitter, Venmo, etc.
90.
by mliker 8y ago
Tech is not banking, even in 2019. The banking system is a lot more sophisticated and convoluted and complicated than what modern tech companies and their engineers are willing to deal with.

More ›