Oleg Ciubotaru
@ociubotaru.bsky.social
Warning: Every post you see on this profile is made by a LLM trained exclusively on fortune cookies with speling mistakes
OpenAI's Random GPT Name Generator Room
April 14, 2025 at 11:41 PM
OpenAI's Random GPT Name Generator Room
Kevin Scott, Microsoft's CTO - thoughts on DeepSeek and Chinese AI companies.
April 1, 2025 at 5:33 AM
Kevin Scott, Microsoft's CTO - thoughts on DeepSeek and Chinese AI companies.
Kevin Scott, Microsoft's CTO, his thoughts on how AI code is following the trend of making it easier for people to talk to machines. You will still need good programmers to build and fix complex stuff...
April 1, 2025 at 5:33 AM
Kevin Scott, Microsoft's CTO, his thoughts on how AI code is following the trend of making it easier for people to talk to machines. You will still need good programmers to build and fix complex stuff...
Some Llama.cpp code is being written by AI.
This is not your typical python hello world app. This is low level C/C++
This is not your typical python hello world app. This is low level C/C++
January 27, 2025 at 6:42 PM
Some Llama.cpp code is being written by AI.
This is not your typical python hello world app. This is low level C/C++
This is not your typical python hello world app. This is low level C/C++
Long but very good read on the recent events in AI: Nvidia Moat, DeepSeek, LLaMa, OpenAI, Anthropic etc
January 27, 2025 at 5:05 AM
Long but very good read on the recent events in AI: Nvidia Moat, DeepSeek, LLaMa, OpenAI, Anthropic etc
Wow :) did I read this correctly?
"DeepSeek’s singular focus on research makes it a dangerous competitor because it is willing to share its breakthroughs rather than protect them for commercial gains."
"DeepSeek’s singular focus on research makes it a dangerous competitor because it is willing to share its breakthroughs rather than protect them for commercial gains."
January 26, 2025 at 8:09 PM
Wow :) did I read this correctly?
"DeepSeek’s singular focus on research makes it a dangerous competitor because it is willing to share its breakthroughs rather than protect them for commercial gains."
"DeepSeek’s singular focus on research makes it a dangerous competitor because it is willing to share its breakthroughs rather than protect them for commercial gains."
Turns out, shipping a solid model is better than having your marketing team blast AI hype headlines on all cylinders.
Less talk, more shipping impressive models!
DeepSeek is now nr 2 free apps.
Let’s see how will OpenAI pricing evolve from here…
Less talk, more shipping impressive models!
DeepSeek is now nr 2 free apps.
Let’s see how will OpenAI pricing evolve from here…
January 26, 2025 at 8:09 PM
Turns out, shipping a solid model is better than having your marketing team blast AI hype headlines on all cylinders.
Less talk, more shipping impressive models!
DeepSeek is now nr 2 free apps.
Let’s see how will OpenAI pricing evolve from here…
Less talk, more shipping impressive models!
DeepSeek is now nr 2 free apps.
Let’s see how will OpenAI pricing evolve from here…
Cheers to the OpenAI engineers being oncall right now.
December 12, 2024 at 12:32 AM
Cheers to the OpenAI engineers being oncall right now.
Also $0.00 - looks great.
I hope this feast of free tokens won't end soon.
I hope this feast of free tokens won't end soon.
December 12, 2024 at 12:06 AM
Also $0.00 - looks great.
I hope this feast of free tokens won't end soon.
I hope this feast of free tokens won't end soon.
Books published before ChatGPT, I see them differently now.
They’re basically backups of organic intelligence... human knowledge, creativity, hallucinations and all.
They’re basically backups of organic intelligence... human knowledge, creativity, hallucinations and all.
December 8, 2024 at 12:13 AM
Books published before ChatGPT, I see them differently now.
They’re basically backups of organic intelligence... human knowledge, creativity, hallucinations and all.
They’re basically backups of organic intelligence... human knowledge, creativity, hallucinations and all.
Don’t use LLMs as your primary doctor, financial advisor, or lawyer - but as a second option - hell yeah!
December 7, 2024 at 3:34 AM
Don’t use LLMs as your primary doctor, financial advisor, or lawyer - but as a second option - hell yeah!
Today I learned that there is a wiki page dedicated to AI doom predictions.
Yann LeCun leads the chart with the P(doom) value of <0.01%
Bottom of the list: Eliezer Yudkowsky & Roman Yampolskiy.
Yann LeCun leads the chart with the P(doom) value of <0.01%
Bottom of the list: Eliezer Yudkowsky & Roman Yampolskiy.
December 4, 2024 at 11:53 PM
Today I learned that there is a wiki page dedicated to AI doom predictions.
Yann LeCun leads the chart with the P(doom) value of <0.01%
Bottom of the list: Eliezer Yudkowsky & Roman Yampolskiy.
Yann LeCun leads the chart with the P(doom) value of <0.01%
Bottom of the list: Eliezer Yudkowsky & Roman Yampolskiy.
Quick update on ChatGPT's impressive growth since its release 2 years ago:
- Hitting 300 million active weekly users
- Over 1 billion messages sent to chatGPT, daily
- 1.3 million developers in the US, with an expanding international user base
- Hitting 300 million active weekly users
- Over 1 billion messages sent to chatGPT, daily
- 1.3 million developers in the US, with an expanding international user base
December 4, 2024 at 11:43 PM
Quick update on ChatGPT's impressive growth since its release 2 years ago:
- Hitting 300 million active weekly users
- Over 1 billion messages sent to chatGPT, daily
- 1.3 million developers in the US, with an expanding international user base
- Hitting 300 million active weekly users
- Over 1 billion messages sent to chatGPT, daily
- 1.3 million developers in the US, with an expanding international user base
Amazon’s new LLMs now take the crown for the “cheapest price for intelligence” that Google coined.
December 4, 2024 at 8:13 PM
Amazon’s new LLMs now take the crown for the “cheapest price for intelligence” that Google coined.
Why would a company that says it's close to achieving AGI need a Chief Marketing Officer?
Wrong answers only.
Wrong answers only.
December 4, 2024 at 12:45 AM
Why would a company that says it's close to achieving AGI need a Chief Marketing Officer?
Wrong answers only.
Wrong answers only.
From the Amazon Nova Family of Models technical report: Just a reminder of how freaking fast Gemini models are. They might not be the smartest, but they're super fast!
Also, strange choice to compare with the older 001 models when the 002 have been out for some time now.
Also, strange choice to compare with the older 001 models when the 002 have been out for some time now.
December 3, 2024 at 11:54 PM
From the Amazon Nova Family of Models technical report: Just a reminder of how freaking fast Gemini models are. They might not be the smartest, but they're super fast!
Also, strange choice to compare with the older 001 models when the 002 have been out for some time now.
Also, strange choice to compare with the older 001 models when the 002 have been out for some time now.
Amazon reminded everyone that they're in the game too.
I'm excited for early 2025: Claude Opus 3.5, Gemini 2, Grok 3, OpenAI full o1, LLAMA 4 [maybe], and seeing Chinese labs pushing open-source models.
I'm excited for early 2025: Claude Opus 3.5, Gemini 2, Grok 3, OpenAI full o1, LLAMA 4 [maybe], and seeing Chinese labs pushing open-source models.
December 3, 2024 at 11:49 PM
Amazon reminded everyone that they're in the game too.
I'm excited for early 2025: Claude Opus 3.5, Gemini 2, Grok 3, OpenAI full o1, LLAMA 4 [maybe], and seeing Chinese labs pushing open-source models.
I'm excited for early 2025: Claude Opus 3.5, Gemini 2, Grok 3, OpenAI full o1, LLAMA 4 [maybe], and seeing Chinese labs pushing open-source models.
Amazon just dropped a gazillion videos from re:Invent 2024. If only there was a way to summarize them... in 1 word.
December 3, 2024 at 11:39 PM
Amazon just dropped a gazillion videos from re:Invent 2024. If only there was a way to summarize them... in 1 word.
December 3, 2024 at 12:27 AM
Tristan Harris: "What I don't think what we're prepared for is we're going to live very soon in a world where 10 years of scientific research could happen in like a single month."
I thought these guys were serious, but this comment easily wins the Bullshit of the Year award.
I thought these guys were serious, but this comment easily wins the Bullshit of the Year award.
December 2, 2024 at 3:43 AM
Tristan Harris: "What I don't think what we're prepared for is we're going to live very soon in a world where 10 years of scientific research could happen in like a single month."
I thought these guys were serious, but this comment easily wins the Bullshit of the Year award.
I thought these guys were serious, but this comment easily wins the Bullshit of the Year award.
I’m not sure how useful it is to automatically store this in ChatGPT’s memory 
December 2, 2024 at 12:33 AM
I’m not sure how useful it is to automatically store this in ChatGPT’s memory 
Fun fact: Horses originally evolved in North America, migrated to Asia, and became isolated there during a warm interglacial period. They eventually died out in North America, only to be reintroduced by the Spaniards in the 1500s.
December 1, 2024 at 11:32 PM
Fun fact: Horses originally evolved in North America, migrated to Asia, and became isolated there during a warm interglacial period. They eventually died out in North America, only to be reintroduced by the Spaniards in the 1500s.
I really liked this part when I was at Meta.
Code [data] wins arguments.
Code [data] wins arguments.
December 1, 2024 at 11:29 PM
I really liked this part when I was at Meta.
Code [data] wins arguments.
Code [data] wins arguments.
OpenAI thinks I talk too much.
December 1, 2024 at 11:28 PM
OpenAI thinks I talk too much.
Try looking up Alexander Hanff too ;)
There is a very good reason for this behavior - and it's called GDPR.
There is a very good reason for this behavior - and it's called GDPR.
December 1, 2024 at 11:02 PM
Try looking up Alexander Hanff too ;)
There is a very good reason for this behavior - and it's called GDPR.
There is a very good reason for this behavior - and it's called GDPR.