Will beatings continue at OpenAI after Sam’s come back? xAI’s Grok launch and its đť•Ź access – is this an ultimate edge? Might be, someday…
Hello Friends!
It’s been a while since the last of those newsletters, but I’m aiming to get more consistent from now on – early New Year’s resolution! I’ll also be providing updates on how my content creator’s journey goes, on top of the Business and Technology topics. Let’s dive right in!
Topics covered this week
My content creator journey update. | AI – Sam Altman firing and hiring. | AI – Grok AI is out and first impressions.
My content creator journey December 2023 update
I’ve been busy in both getting to know things as well as creating. Having initial success on 𝕏 with a post going viral to get 200k impressions I embarked on a journey to get 5 milion impressions within three months and getting monetized. You can follow the progress weekly in this 𝕏 thread.
Putting and getting impressions there allowed me to meet people, which is awesome. I also had a peek preview of new AI tool, Freepik Pikaso, more on which I’ll share soon.
This week I took part in two IT events in the local community, which helps me stay in touch with the pulse of the business. These were QA / Software Testing event organized by local community group LuczniczQA and chapter meeting of The Association of Business Service Leaders (ABSL)
Focusing on đť•Ź made me spend less time on other platforms – LinkedIn, YouTube, Instagram. I’m also in a trial-and-error phase if I should post under Business and Technology brand profiles or using my personal ones. Testing the waters with both and data will lead the decisions.
Sam Altman firing and hiring
This was better than any other soap opera I’ve ever witnessed, with dopamine hits after news unfolding in real time, one after another.
The events in chronological order:
Nov 17 (Friday): Sam Altman got fired from OpenAI. The board lost confidence in him. Mira Murati appointed interim CEO.
Nov 18-19: negotiations take place, Sam Altman back at OpenAI HQ with Guest_04 pass instead of his own ID. Seems like he might be back within hours.
Nov 20 (Monday): Sam Altman is not coming back to OpenAI. After weekend of negotiations board appoints Emmett Shear, former Twitch CEO, as new interim CEO of OpenAI. Mira Murati supposedly wanted Sam back, Ilya Sutskever supposedly said it’s not an option.
Satya Nadella, Microsoft’s CEO announces that: “we’re extremely excited to share the news that Sam Altman and Greg Brockman, together with colleagues, will be joining Microsoft to lead a new advanced AI research team”.
Ilya Sutskever (OpenAI founder and a board member at the time): I deeply regret my participation in the board’s actions. I never intended to harm OpenAI. I love everything we’ve built together and I will do everything I can to reunite the company.
Almost all of OpenAI employees, including Ilya, sign a petition to bring Sam back.
Nov 21: Current state of affairs seems everyone except the board questions decision to fire Sam and he’s still in the game to be back as OpenAI CEO. Emmett is asking for the real reason to fire Sam, or he’ll quit / not take the role at all.
Nov 22: OpenAI’s announcement: We have reached an agreement in principle for Sam Altman to return to OpenAI as CEO with a new initial board (…).
OpenAI’s structure with a nonprofit trying to have an oversight over for-profit LLC is not working, and it cannot be working due to the number of entities there and how those are intertwined.
Sam being back with huge backing of the employees threatening to leave gives him a lot of power. At this point I am disillusioned with any oversight being in place.
Why have employees backed Sam Altman? On one hand he’s successful building products and making deals, which increases OpenAI’s value, directly affecting their compensation and future prospects. On the other he might have been a good boss. But has he?
Some light has been shed on it by Ilya Sutskever, who posted:
I learned many lessons this past month. One such lesson is that the phrase “the beatings will continue until morale improves” applies more often than it has right to.
Post has been quickly removed from the Internet, but screenshots remain. Was the cause Sam’s retaliating against Ilya? Time might tell, but for now it’s pure speculation.
Google Meet vs Microsoft Teams is a fun spin on the OpenAI leadership firing story. Expulsion took place through Google Meet. Once Satya posted that they would join Microsoft it was taken that now they will need to use Microsoft Teams going forward.
I wasn’t sure why people make fun of Microsoft Teams, I thought it is as bad as Google Meet. Recently though I was forced to use Teams after years and it’s simply too heavy, too slow and crashes too often.
Grok AI is out
x.AI announced Grok – their first large language model. The announcement made on November 5th has two statements I find profoundly important:
(…) A unique and fundamental advantage of Grok is that it has real-time knowledge of the world via the 𝕏 platform. (…)
Grok is still a very early beta product – the best we could do with 2 months of training – so expect it to improve rapidly with each passing week with your help.
Data is key in all the language models. It actually might be the only differentiator between those in the long term. đť•Ź platform, formerly Twitter, is one of the best places to get real time information about what is going on in the world. Having exclusive access to content there and being able to use nearly in real time is a huge advantage over other language model providers.
This clearly shows in one of the comparisons I’ve seen after the announcements, between Grok and Phind Model. I’ve thrown Bing search, ChatGPT 3.5 and Bard answers and only Google’s product provided accurate answer, along with Grok. Real world, current, data access is key.
Even with 𝕏 platform or the Internet access, it is not easy to distinguish good text from all of the clutter. Signal to noise ratio is not particularly good. More on that in Elon Musk’s latest interview with Lex Fridman (link below).
Comparison of Grok, Phind, Bing and Bard, ChatGPT 3.5 answers regarding Joe Rogan’s latest interview:
Grok
Phind:
Bing:
Bard:
ChatGPT (GPT-3.5):
Secondly, the announcement x.AI said that it took two months to train the model. Such short development cycle and the fact the tool was opened for testing is the way all Elon Musk’s companies operate.
As of now Grok is rolling out to 𝕏 premium+ subscribers in the US. I’ve subscribed and used some VPN tricks to get hold of it and my initial takeaways are:
- Grok’s access to 𝕏 post is not real time, but it’s fairly robust. Posts from a day or so ago are used in the responses, also links to posts are provided at the end of the answer.
- I have managed to use it for some brainstorming, similarly to OpenAI’s ChatGPT (3.5). Some examples of that is searching for a word to be used in the text:
- Grok’s context window seems to be the size of a goldfish. This is not enough to have decent conversation. We’ll definitely need Grok 2.0 with bigger context window.
Coming up soon
Google’s Gemini, Freepik Pikaso, Cybertruck’s 48V and other features
My Weekly Media Roundup.
Full x.AI announcement about Grok on đť•Ź platform
https://twitter.com/xai/status/1721027348970238035
The All-In Podcast E157: Epic legal win, OpenAI’s news deal, FCC targets Elon, the limits of free speech & more
https://www.youtube.com/watch?v=w85An9FVsb0&t=1s
Deep Dive with Ali Abdaal – How Writing Online Will Make You A Millionaire – Nathan Barry
https://www.youtube.com/watch?v=LPkNt1gFdZ4
Ali Abdaal – Everything’s Changing – Life Update https://www.youtube.com/watch?v=XtOshDbMGkA
Munro Live – Elon Musk and Sandy Munro discuss the Cybertruck!
https://www.youtube.com/watch?v=ky1Z2klPalw
Google: Hands-on with Gemini: Interacting with multimodal AI
https://www.youtube.com/watch?v=UIZAiXYceBI
This Business And Technology blog newsletter is curated by Ariel Zwolinski
You can find the author at: x.com/arielzwolinski | linkedin.com/in/zwolinski | Substack