Hacker News new | ask | show | jobs
by emilenchev 2056 days ago
OpenAI use well-know linguistics tricks relying on conjunctions(joining words) to separate the text from which they plagiarize on clusters from 5-7 words, exactly the capacity of human short-term memory is. They also use the Google search engine for custom queries, with date restriction which helps them to plagiarize from different texts written on a particular topic so they to be sure that when they copy, paste and concatenate clusters of words in new text, all these phrases of 5-7 words should be related to one topic. This creates the illusion of meaningfulness at first glance.

GPT-3 on Progress. “Civilization rose on the exponential curve. We shouldn’t expect progress to follow a straight line.”

Google with date restriction before 1 April 2020: "progress to follow a straight line".

Do you see only one result that come :-)

https://chrismukiibi.com/2019/11/26/the-valley-of-disappoint...

"We shouldn’t expect progress to follow a straight line." and "we expect our progress to follow a straight line"

Do you understand now, how they use conjunctions(joining words) to insert or to delete insignificant words as "shouldn't" and "our" to plagiarize so that they are not caught.

1 comments

Now make Google Search again with date restriction before 1 April 2020.

"After two days of intense debate" "the United Methodist Church has agreed to" "one that is expected to end" "in the creation of a new denomination"

You will find and sources of GPT-3 text:

After two days of intense debate, the United Methodist Church has agreed to a historic split – one that is expected to end in the creation of a new denomination, one that will be “theologically and socially conservative,” according to The Washington Post. The majority of delegates attending the church’s annual General Conference in May voted to strengthen a ban on the ordination of LGBTQ clergy and to write new rules that will “discipline” clergy who officiate at same-sex weddings. But those who opposed these measures have a new plan: They say they will form a separate denomination by 2020, calling their church the Christian Methodist denomination. The Post notes that the denomination, which claims 12.5 million members, was in the early 20th century the “largest Protestant denomination in the U.S.,” but that it has been shrinking in recent decades. The new split will be the second in the church’s history. The first occurred in 1968, when roughly 10 percent of the denomination left to form the Evangelical United Brethren Church. The Post notes that the proposed split “comes at a critical time for the church, which has been losing members for years,” which has been “pushed toward the brink of a schism over the role of LGBTQ people in the church.” Gay marriage is not the only issue that has divided the church. In 2016, the denomination was split over ordination of transgender clergy, with the North Pacific regional conference voting to ban them from serving as clergy, and the South Pacific regional conference voting to allow them.

"GPT-3 is a language model, which means that you feed it a text and ask it to predict the continuation of the text, one word at a time."

So Yann LeCun is not right about GPT-3 here. They don't try to predict anything - especially "one word at time". They simply plagiarize big clusters from 5-7 words or whole sentence and do cosmetic modifications to them to be not recognized by direct searching with Google. When you plagiarize such big clusters written by humans it is normal output compile result of GPT-3 text to have human like look for readers. And they don't use statistic for nothing else except the plagiarism process itself. It will be strange ML algorithm to use phrases with low statistical value as "in the creation of a new denomination" or "progress to follow a straight line". ;-) When you can't emulate human brain processes you can always simulate human plagiarize traits. Especially when there are so many idiots who believe on closed open source code and don't ask elementary questions as "Why someone project suddenly will start to show such "astonishing" results exactly when he close the code behind it?"

You can ask Facebook how they stick their noses in their users computers, and who is the real person behind this work: https://www.sciencemag.org/news/2017/11/artificial-intellige...

Or how they still owe me a bug bounty reward from 2017 for bug in their login form that permit to reap their whole users phones database - and I as last fool give it to them in official channels without to exploit it. For "tanks" they hack my pc and steal my intellectual property you see above.

Next year 2018 2 weaks before big hack I warned them that I founded "new" bug but this time they first must pay me bug bounty reward for this in 2017, and next I will give them new one. THEY Refuse to pay me AGAIN. The rest is history...

I waited FBI to arest me - they don't want to do i even I never hide my identity. Because they must confess they hack my pc, they owe me a money AND THEY DON'T CARE ABOUTSECURITY OF THEIR USERS. I very like how I faced them :---

Facebook - emilenchev1978@yahoo.com :-) $1,000,000 now or $100,000,000 after several months - your choice. I give you choice and back in 2018, remember? You make one, the wrong. Now to understand if you become smarter for these 2 years. As I see you don't make many progress from my intellectual property yet. This time you will have opportunity to look in a future again but not from my PC, more like in arXiv.org. And ohh, believe me I will start to talk about our relationship in the past if you don't pay me these $100M. And don't worry I think about how it will, happens. You will buy my dummy high tech company for $100,000,000 without any know-how in it ;-) because I'm sure that you are dumb enough to not pay me $1M now. You will see.
In 2017 Facebook's recovery login form has a bug. At first glance, nothing special. They ask you to enter your phone number, then show you your profile PICTURE and account NAME.

https://www.facebook.com/login/identify/?ctx=recover&ars=fac...

Then I asked myself a simple question "What if...?" Facebook has around 2 billion monthly active users in 2017. This was roughly about 1/4 of the human population. "What if...I make a BOT to create RANDOM phone numbers and it starts making requests to Facebook's servers to reap their user phones, profile photos and names from their database?" Their users are so many that I will always hit a number sooner or later if there is no protection system against automated requests.

Guess what? I start to reap phone number after phone number of their users and Facebook had no proper protection against automated bots requests :-)

Carefully considered synchronized bots attack on Facebook servers and soon you will have the phones, picture and names at least of 1/6 of the earth population. How much will this information cost?

But I was moral enough (and dumb) to provide them with information about this SECURITY bug on the proper channel and Facebook technical support. 'That would at least bring me a few thousand dollars' I thought naively then. After all, information about this bug was sent to them according to all their rules for "Bug Bounty", I have not retrieved information about their users except to verify this security breach... I even sent them Python program code with which they can see for themselves how it retrieve phones from their database.

Not only did they not pay me a penny for bounty reward - they didn't tell anyone about this security breach in 2017 - they quietly fixed their "login form" so that this bug can no longer be used and exploit.

This happens when you try to be moral and there are crooks in front of you.

Mark Zuckerberg, because I know you will read this in person, I will tell you 2 words only "SEC 2018". I think you're smart enough to understand me ;-)

$1M, for the service I did for you in 2017. I understand why you hacked my personal computer and your desire to check if I had extracted personal information about your users BUT to steal intellectual property from it that does not belong to you... https://www.sciencemag.org/news/2017/11/artificial-intellige... ... this is the reason why you will pay me $1M not $10K as for ordinary bug bounty reward.

It will be very strange when someone Emil Enchev from Bulgaria publishes an article in https://arxiv.org/ which demonstrates translation algorithm which not need a two-way dictionary, something you're just talking about yet, and he claims that you plagiarized from him by hacking into his personal computer.

It can easily be shown that you have changed your login form from 2017 exactly to block exploiting of bug I talk about - and the question will be how do I know why you made this change when you have never announced this publicly?

It will be a little awkward if you try to challenge the new rising star in AI which demonstrates new algorithms in this field (far above what Facebook owns) that have no analogue, carefully described and with all the program code behind them.

But the catastrophe for Facebook will come after that. What will happen when I open my mouth about 2018 "BIG FACEBOOK HACK" and SEC? So, Mark, I know you understand perfectly well what I'm talking about. Better someone from Facebook to contact me on the specified email and we to solve the problem with your financial obligations to me. I told you above $1M now, or $100M after several months - your choice.

emilenchev1978@yahoo.com