XGCAST

Predictive Asset Intelligence

[XGCAST @xgcast_]

Escritorios desordenados, mentes brillantes.
-Steve Jobs en 2004, fotografiado por Diana Walker.
-Elon Musk en 1994 o 1995, fotografiado por Jennifer Gwynne.
-Jeff Bezos en 1999, entrevistado por 60 Minutes.
-Dieter Rams en Braun, 1970. Fotografiado por Abisag Tüllmann.

[XGCAST @xgcast_]

How it started...

[Sam Altman @sama]

today we launched ChatGPT. try talking with it here:

https://t.co/uWra8LKFMN

[XGCAST @xgcast_]

@thsottiaux Great News!

[XGCAST @xgcast_]

Wake up, Neo...■

[XGCAST @xgcast_]

@convolutiva @OpenAI Esto todavia es muy cierto...

[XGCAST @xgcast_]

@JoviNomas @Megadeth @OzzyOsbourne Y para los que quieran una guitarra de verdad!
https://t.co/aRwQwKDh6n

[XGCAST @xgcast_]

@rockaxisoficial Para los que quieran una guitarra de verdad!

https://t.co/aRwQwKDh6n

[XGCAST @xgcast_]

@ChrischileXD Acá un súper buen artículo de por qué fracasan los emprendimientos:
https://t.co/sKbv7FKMco

[Interesting AF @interesting_aIl]

Fascinating visual representation of morse code

[XGCAST @xgcast_]

@SnazzyLabs To anyone, 24 or 32g?

[XGCAST @xgcast_]

@ahorrandoclick1 Amo tus publicaciones ahorrandoclicks!

[XGCAST @xgcast_]

@miniapeur It makes the pain go away…

[XGCAST @xgcast_]

@IMFDChile @CNNChile @dccuc @cen_ia @CentrosAnid

[XGCAST @xgcast_]

@MacRumors New model comming?

[XGCAST @xgcast_]

@LinusEkenstam For your spanish audience.
https://t.co/uwFrDhZzQM

[XGCAST @xgcast_]

@EntelExploralab @EntelDigital Hola @EntelExploralab ! Donde publican su agenda de actividades?

[XGCAST @xgcast_]

@natkohasic Your work is awesome. Best i ever seen. Keep doing beautiful things! Cheers from Chile!

[XGCAST @xgcast_]

@robertocastro Totalmente conectado con este video! “¿Estas dispuesto a correr con todo si la distancias es desconocida?”
https://t.co/w7TMWKrITo

[Sahil Bloom @SahilBloom]

I can't stop thinking about this question:

Are you willing to sprint when the distance is unknown?

In 2021, Georgia Tech strength coach Lewis Caralla delivered this epic speech to the football team.

If it doesn't get you motivated, you may need to check your pulse...

He opens with a few harsh truths:

• Winning isn't loyal to you
• Winning doesn't care about you
• Winning doesn't care how sore you are
• Winning doesn't care how hard you work
• Winning doesn't care how much sleep you get

But it's his question that stuck with me:

Are you willing to sprint when the distance is unknown?

The willingness to sprint with no clear view of the finish line is rare.

It requires two things:

1. A deep belief in one's self
2. A deep belief in the mission

If you have 1 but not 2, you won't be able to do it.

If you have 2 but not 1, you won't be able to do it.

You need both.

In my observation, the greatest things in life are accomplished when you're willing to sprint when the distance is unknown:

The sprint to care for your loved ones in their time of need.

The sprint to build something meaningful.

The sprint to serve others and create positive ripples in the world.

Goal: Find those rare things in life that you're willing to sprint for when the distance is unknown.

That, to me, is the definition of winning.

"And why chase winning? Because the only thing that's guaranteed in life if you don't chase it is losing."

[XGCAST @xgcast_]

@convolutiva @bindureddy In Chile, the legislator incorporated statistical methods into the definition of AI, which would imply asking the State for permission to even calculate P(a|b). More, we must register users who consume the function. Can you believe it?
https://t.co/9PpBrc4Vdm

[Bindu Reddy @bindureddy]

Regulating AI Severely Limits Open-Source AI

Some folks including Elon Musk hold the position that you should aggressively regulate AI while at the same time being pro open-source.

Unfortunately, calling for aggressively and pre-maturely regulating AI will likely severely limit open-source AI.

Today, when there is NO REGULATION, open-sourcing software, including an LLM in a large organization like Meta, Google, or Amazon, is highly non-trivial.

The innovating team that is open-sourcing the software has to jump through many hoops today - Several tickets need to be cut, approvals obtained and naysayers silenced.

The PR and comms teams have to be convinced that there won't be any negative press and the compliance teams that there is little to no chance of additional regulatory or government scrutiny.

Now imagine what happens when there is AI regulation under the pretext of "national security."

The number of tickets, approvals, and naysayers are going to explode. The open-sourcing team will be putting their careers in these large companies on the line to go out on a limb and fight for open AI.

The number of people willing to take this risk will go down severely.

Now imagine if there is some team willing to take this risk and who succeeds in open-sourcing an LLM.

Now, if by chance, that LLM is used to generate misinformation or something that can even be remotely considered harmful to national security or if a vulnerability is exposed in the LLM that violates this regulation, then the organization will have to face negative consequences.

This will inevitably cause the comms and compliance teams to dig in and shut open-sourcing down!

In essence, you don't have to explicitly ban open-source; you can indirectly limit it severely by introducing regulation.

Sure, open-source companies like Mistral may still open-source AI, but that isn't the same as being GPU and data-rich and having big tech companies like Meta being able to open-source LLMs!

TLDR: regulating AI will slow down and limit the release of open-source models, especially from larger organizations.

The corporate lobbyists will succeed in regulatory capture and protecting their dominance and investment unless a big tech like Meta commits to open-source even after AI is regulated.

[XGCAST @xgcast_]

@JonErlichman Yes, but the point here is not a confrontation with Microsoft, but to clarify ideas by writting.