Those were the Good Old Days of Computer Networking

[RE]

I remember when we got our first Ethernet setup in the common area on the ground floor of each of the dorms on the main campus at Columbia in 1977.  Prior to that the only network with printers attached you could send you word processor produced papers to for printing were up in the Computer Lab, brand new when I arrived in 1974 for my Freshman Year.  After my HS years working with a Brother electric typewriter which was a huge step up from the Royal manual typewriter I had in JHS, the Word Processor/Printer combo with the ability to correct your typos, change fonts, use bold and italics was like magic.  No more Whiteout, no more corrasable paper with smudges, no more carbon paper. It was amazing. 

Our dorm network had a primitive all-text messaging system, and the OP system did not yet have a Point-Click Graphical User Interface (GUI) and used the MS-DOS OP system Bill Gates made his initial fortune on before Windows, which was a direct ripoff of a XEROX GUI.  I had my very first email address, and our network was connected to other university networks via the Arpanet, before the World Wide Web.

It kept getting better through about the mid 1990s, when both Microsoft & Apple started going ballistic with Digital Copy Protection, and one by one all the free and open sources of information started getting locked up tighter than a 13 year old Catholic Virgin with a Kevlar Chastity Belt inside the Chapel of the NORAD command center underneath Cheyenne Mountain.  Rolling out around the same time was Google's all seeing, all knowing data collection and search cancer, metasticizing its tentacles through every aspect of people's lives, so marketers can know precisely how many squares of toilet paper you use every time you take a shit.

Once upon a time though, there was Computer Camelot.

https://spectrum.ieee.org/ethernet-ieee-milestone

Ethernet is Still Going Strong After 50 Years

The technology has become the standard LAN worldwide

RE

[K-Dog]

The good old days.  ASCII text porn wasn't moving in 1977 ................................. but you were!



The days when the Doomstead was a vagabond at GLOBAL COLLAPSE waiting for cheap server space might not have allowed this.

Once upon a time in .

[RE]

The good old days.  ASCII text porn wasn't moving in 1977 ................................. but you were!

Yes, me and Fred Flinstone were moving in those days







RE

[RE]

Unfortunately, I have used up my last free article for the month from behind the New Yorker paywall, so I haven't read the article.  I figured it might interest K-Dog tho.

https://www.newyorker.com/magazine/2023/11/20/a-coder-considers-the-waning-days-of-the-craft

A Coder considers the waning days of the Craft

RE

[K-Dog]

Unfortunately, I have used up my last free article for the month from behind the New Yorker paywall, so I haven't read the article.  I figured it might interest K-Dog tho.

https://www.newyorker.com/magazine/2023/11/20/a-coder-considers-the-waning-days-of-the-craft

A Coder considers the waning days of the Craft

RE

Thank you it is a good article.

It is about coding and how GPT-4 can cough up code on demand.  The contention is that the programmer is as out of work as Ned Ludd.  Sort of because in the end the author writes:

Computing is not yet overcome. GPT-4 is impressive, but a layperson can't wield it the way a programmer can. I still feel secure in my profession. In fact, I feel somewhat more secure than before.

Besides agreeing with the author, it seems I can one-up the professionals.  I can get results from GPT-3 --- they apparently need GPT-4.

GPT-3 helps me code but only because I know what the fuck I am doing, and because I can tell it to fuck off any time I want.  I catch as many mistakes it makes, as it catches of mine.  Lately I catch more!  The time it takes to tell it to do something can take as long as doing it myself.  The biggest payoff is that A.I. has given me more confidence in my own abilities.

GPT-3 is an idiot savant.  Stupid, but with vast knowledge.  If you can ask the right question, hours of searching Stack Overflow are avoided.  Stack Overflow was the old way of finding what you needed to know when you got stuck and needed the secret sauce.  Stack Overflow is a network of programmers who ask and answer questions.  I have not contributed there much, but my reputation is really good.

A. I. is a road on which programmers may drive.  But if you have to go cross country, a coder has to hike alone because A.I. will faceplant on the first tree.

* This thing about programmers being put out of work?  Didn't B. Gates start flying coders in by the thousands about thirty years ago because there were no American coders to be had?

Yes, that bullshit flew and NOBODY cared. 

Consequently software employment is a bit of a rigged game as it is now.  So A.I. making a pig dirty?  I do not think so.  Many talented people have been unemployed in America for a very long time.  A.I. did not make that happen.  A.I did not screw them.

People made that happen.  People screwed them.

Swagath is 150 feet west of where I work in a warehouse doing my undercover proletarian investigative work.  I was investigating today.  It is one of at least 100 Indian Businesses within a mile of work.  I am NOT exaggerating. 

You are looking north.  I park next to the pet clinic.  And I do like curry.

I can code.  I can also throw a sixty pound box onto a shelf that is seven feet high.  Not a small feat for someone my age.  Proletarian undercover work is not without its rewards.

Thanks Bill.  My heart will flutter with gratitude when I throw the next box up on the shelf.  And thank you A.I. for showing me I really do have mojo.

[Knarf]

  We attracted many a "chela" when the newsgroups on usenet was kicking. alt/buddhism...alt/zen...etc. That was in the early 90's.
  I just read an article that I am going to post in full because I want know the real impact of the GPT programming and it's explosion into EVERYTHING! If this article is close, then we are in for some enormous "human" adaptation...fast.

OpenAI Just Killed an Entire Market in 45 Minutes
The Story Everyone Should Have Seen Coming
Ignacio de Gregorio
Ignacio de Gregorio

·
Follow

6 min read
·
Nov 9

This is a story about how a 45-minute speech killed an entire billion-dollar market.

This week OpenAI had its first-ever developer conference. Fewer times in history had there been higher expectations of a private technology event, and the OpenAI did not disappoint.

They announced a new, smarter GPT-4 model named 'Turbo' and they included several fascinating features that severely improve outputs and offer developers much more control over their models.

And all this while still managing to decrease, once again, API costs.

But one announcement went swiftly past most of the media while being the one with the highest immediate impact on the AI market, as this new feature alone threatens an entire start-up ecosystem and billions of invested capital.

Most insights I share in Medium have previously been shared in my weekly newsletter, TheTechOasis.

If you want to be up-to-date with the frenetic world of AI while also feeling inspired to take action or, at the very least, to be well-prepared for the future ahead of us, this is for you.

🏝Subscribe below🏝 to become an AI leader among your peers and receive content not present in any other platform, including Medium:

Subscribe | TheTechOasis
The newsletter to stay ahead of the curve in AI
thetechoasis.beehiiv.com

The Importance of Abstraction
Even OpenAI couldn't have predicted it. Not even they saw ChatGPT coming.

When you even impress yourself
Over the last few months, we have seen multiple accounts from the top management on OpenAI like Ilya Sutskever, Mina Murati, and Sam Altman claim that when they released the open beta for ChatGPT back in November 2022 they didn't expect it to have that huge access.

In two months, it had already reached 100 million first-time users, and was already officially the fastest-growing product in history, only recently surpassed by Instagram's Threads.


Source: ExplodingTopics
ChatGPT has now crossed the 100 million active monthly users mark.

In Ilya's case, he went as far as claiming that he honestly thought the product 'wasn't that good'.

Nowadays, few people in this world don't know what that 7-letter name means and many see Generative AI not only as a new frontier for AI but even as a new computer paradigm.

But even though the product was amazing, you really couldn't do much more than simply toy with it. It was a fun and bewildering thing to use and... that was pretty much it.

To use it for real, valuable use cases like a knowledge management system — far and wide the most common enterprise Generative AI use case today — you required additional tools like vector databases.

Also, when OpenAI released its first ChatGPT API, people soon realized how constrained the browser version was compared to the API one.

It had a much smaller context window (the amount of words it can process for one request) and the opportunities for a more customized solution were almost non-existent.

Long story short, unless you were a seasoned developer, ChatGPT was nothing more than a pastime product with some minor utility if you were willing to put on the effort and you had some tech savviness.

And with this realization, opportunity came.

Millionaire 18-year-olds
Soon, multiple developers around the world realized that they could create entirely new products in a few days that could offer more customized solutions built on top of ChatGPT by abstracting the complexity of building such solutions to more layperson audiences.

For instance, tools like ChatPDF or AskYourPDF allowed you to send the model actual documents for it to answer about them, and so became widely available.

As the appeal was very clear, people started paying decent money to use these products even though ChatGPT was (and still is) mainly free.

The name of these types of start-ups?

ChatGPT Wrappers, and soon enough reports of 18-year-old bedroom developers around the world earning, in some cases, millions, proliferated.

Probably, the most famous of all these wrappers is Jasper, a ChatGPT-based product focused on marketing teams.

The tool allows you to create copyrighting material at scale, write blogs, optimize your web pages for SEO, and many other features.

Despite basically being a digital user interface with marketing-specific prompt engineering on top of OpenAI's ChatGPT APIs, the company still managed to raise $125 million at a staggering $1.5 billion valuation back in October 2022.

Considering that ChatGPT was released literally the next month, that was probably one of the worst-timed investment rounds ever.

But, after OpenAI's developer conference, this ecosystem of start-ups being built on ChatGPT will soon look more like a war zone.

Jasper And The Wrapper Graveyard
Long story short, OpenAI appears to be focusing its strategy for growth on making building AI tools easy by vertically integrating their solutions into higher abstraction layers... where the wrappers happen to be making their bucks.

Building time machines
Honestly, it was quite easy to predict because, as Scott Galloway has said many times, the easiest way to build a billion-dollar company is to build a 'time machine'.

A product or solution that makes the lives of people who use it easier.

More productive. Less effort.

In the case of OpenAI, that is none other than making Generative AI, a technology that a few years ago required almost a doctorate to even understand, an easy-to-use tech that works by simply clicking a few buttons... or simply asking something to it.

As shown by Sam Altman in OpenAI's dev conference, a new software-building paradigm is upon us, with examples like Canva GPT where you can build Canva posters by simply asking for it.

This may seem eerily similar to 'ChatGPT plugins'. In fact, Sam itself mentioned that these GPTs are indeed an improved evolution of that concept.


Source: OpenAI
In other words, they are making AI declarative.

Declarative developers
Just like you can find imperative programming languages like Python, where you have to control the complete building process, and declarative programming languages like SQL, where you simply declare what you want and the script gives it to you, AI is moving from an imperative technology to a declarative one.

And with the new addition to an already overwhelming ChatGPT API of the Assistant API, you will now be able to embed agents into your solutions that leverage ChatGPT to use tools and do the dirty work for you.

This feature, added to the GPT Store where developers can build customized solutions and share revenues with OpenAI, the improved function calling, JSON outputting, seed referencing, accessible log probs, and more capable but cheaper models like GPT-4 Turbo, represent new amazing opportunities for developers to build incredible solutions.

The dawn of a new computing platform
Previously to this announcement, you required these wrapper products to do the most simple stuff.

Now, most of their features are out-of-the-box capabilities in ChatGPT.

OpenAI is slowly pulling everyone into its ecosystem with a carefully laid out strategy that makes it almost indispensable to be a part of its platform if you want to survive.

And just like before, a new ecosystem of applications will be built on top of ChatGPT, but this time is different.

This time, to play the game you need to play by the rules of OpenAI, and that is a death sentence to many start-ups and the billions invested in them.

Dear Jasper, it was a very fun... but short ride.