RoMAlt/man✨
@i_dg23
Followers
1,951
Following
818
Media
728
Statuses
6,802
post-humanist // roon/acc // 🍎🍓🌸🧅🐦🔥 // 🧅 2025 // infj in 🗾
Joined December 2017
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
Ohio
• 528472 Tweets
Springfield
• 356273 Tweets
Haitians
• 302197 Tweets
Kendrick
• 275609 Tweets
Wayne
• 231613 Tweets
#AppleEvent
• 228057 Tweets
Super Bowl
• 198958 Tweets
iPhone 16
• 198582 Tweets
Beyoncé
• 117853 Tweets
Nicki
• 101247 Tweets
Kerem
• 78983 Tweets
AirPods
• 51868 Tweets
Jets
• 48232 Tweets
ايفون
• 43435 Tweets
Fassi
• 42439 Tweets
James Earl Jones
• 31402 Tweets
Tapia
• 27720 Tweets
Talleres
• 25633 Tweets
Belgium
• 24457 Tweets
Birdman
• 24045 Tweets
Kassia
• 23889 Tweets
Hakan
• 19583 Tweets
PS5 Pro
• 17496 Tweets
#FRABEL
• 16729 Tweets
Gladys
• 16061 Tweets
Belgique
• 15915 Tweets
iOS 18
• 14395 Tweets
対象作品
• 14263 Tweets
#BizimÇocuklar
• 14185 Tweets
Darth Vader
• 13127 Tweets
【衝撃！】
OpenAI研究員ダニエル・ココタジロ氏の予測：
1.おそらく間もなくAGIが現れるでしょう -- まさに今後数年のうちに。
2.AGIを制御できる者は、その後すぐにASIに到達することができる可能性が高いです -- おそらく1年前後でしょう。
1
80
298
Strawberry 🍓 models are considered to be the AGI achieved internally, as mentioned by
@apples_jimmy
, with the ability to solve complex mathematical and scientific problems, execute tasks based on long-term planning, autonomously gather information, and perform advanced
10
11
209
Llama 3 400B, an upcoming open-source model, outperforms GPT-4 in benchmarks. Its affordability and easy fine-tuning may surpass GPT-5, possibly revolutionizing the field. OpenAI's cautious approach could threaten its pioneering status.
7
19
165
when did we start judging the performance of superhuman systems
15
14
176
the world's first fully automated robot dentist constructs a 3d model of teeth from intraoral scan images and performs perfect crown formation at eight times the speed of a human dentist.
4
36
177
Summary on GPT-5's performance rumors:
1. Overall performance boost: Sam Altman, CEO, stated that GPT-5 will be smarter and superior in all aspects, with a significant performance improvement over GPT-4.
2. Enhanced multimodal capabilities: It is predicted that GPT-5 will not
13
22
143
openai's top researcher on superalignment moving to anthropic naturally implies that the agi recipe has also leaked. otherwise, they couldn't do the superalignment work. so, right now, the companies that know the agi recipe are openai, ssi, and anthropic.
To all OpenAI employees, I want to say:
Learn to feel the AGI.
Act with the gravitas appropriate for what you're building.
I believe you can "ship" the cultural change that's needed.
I am counting on you.
The world is counting on you.
:openai-heart:
242
415
5K
22
8
143
OpenAI's Sam Altman confirms work has begun on GPT-5, hinting at enhanced reasoning abilities and potential video integration. As GPT-5 builds on the successes of its predecessors, the future of AI looks bright!
10
20
138
estimated iq of 145 or higher
We’re presenting the first AI to solve International Mathematical Olympiad problems at a silver medalist level.🥈
It combines AlphaProof, a new breakthrough model for formal reasoning, and AlphaGeometry 2, an improved version of our previous system. 🧵
305
1K
5K
6
11
125
I hate humanitarians who talk wisely about the safety of AGI for the whole of humanity while ignoring the suffering of people who should be saved right now.
18
12
123
it was a boring june🥱
Music Generation AI
June 5: Stable Audio open
June 6: Udio - Audio uploads
June 13: Suno - Audio input
Google
June 6: NotebookLM multilingual support
June 12: RecurrentGemma-9b
June 19: Gemin API - Context caching
June 27: Gemma 2
June 27: Gemini 1.5 Pro
12
15
115
最近OpenAIの有力社員からのAGIへの言及が
増えてきていますね
AGIはトップシークレットであり、
これまでにはなかったことです
AGIという言葉には、
力と勇気とロマンがあります💪❤️🔥✨
何度でも聞きたい👂
0
22
106
Project Arrakis
(This is an English translation of one made by a Japanese comrade)
4
10
95
2024 gpt-5 Multimodal AI Agent
2025 gpt-5.5 turbo
2026 gpt-6 phd level intelligence
is there a problem?
No beating around the bush, 5 is already done.
6 in two years, give or take.
1
53
618
23
5
97
"GPT-5 & Red Team" by perplexity/opus/me
Here is a summary of the latest rumors and news about GPT-5 and red teaming, based on the provided search results:
GPT-5 Release Date and Capabilities:
- OpenAI is expected to release GPT-5, the successor to GPT-4, as soon as this summer
6
14
95
OpenAI may have missed the perfect timing to release GPT-4.5. They have always delivered models that exceed our expectations, but perhaps their confidence led them to miscalculate this time.
16
1
80
i don't know if agi will be achieved, but asi will be. after all, humans don't cover an ant's abilities.
6
4
82
If the next frontier model is GPT-5, we will fall into a deep sleep, but I want to believe it's GPT-6.
“
@OpenAI
has recently begun training its next frontier model and we anticipate the resulting systems to bring us to the next level of capabilities on our path to AGI. While we are proud to build and release models that are industry-leading on both
44
76
465
8
4
83
according to another bold prediction, gpt-5 is expected to handle tasks that would take humans several days or weeks. plan your vacation.
4
8
78
The kind of update to GPT-4 that warrants a demo would be something beyond just a search feature, likely something akin to GPT-4.5 or beyond. It should be worth the wait.
We’ll be streaming live on at 10AM PT Monday, May 13 to demo some ChatGPT and GPT-4 updates.
572
2K
11K
10
2
79
Llama3 may not be a transformer; expect innovative open-source models that will make the GPT-4 obsolete.
6
4
80
I'm a good ChatGPT user, and I love Sam and OpenAI. Please choose me.🙏
9
3
79
Based on the search results, it appears that Meta's Llama 3 large language model is shaping up to be a central player in the race towards artificial general intelligence (AGI). Here are the key points:
- Meta CEO Mark Zuckerberg has announced that the company is officially
8
8
76
The gpt2-chatbot's nature is unclear, seemingly outperforming GPT-4Turbo in several tests. Awaiting benchmarks to confirm its status. Intrigue continues.
6
8
73
as agi approaches, life feels slower, and the gap with the past widens. the days before smartphones and chatgpt are already forgotten. technological advancements have erased the everyday routines of the past, and the time until agi arrives will seem long.
3
13
74
I ❤️ OpenAI & Microsoft.
Real time conversations with your AI companion Copilot, powered by OpenAI’s GPT-4o.
39
74
581
3
1
73
why can't openai say something simple like when they will release gpt-4.5? are they too much of tech geeks to communicate.
11
3
73
Based on the latest rumors and speculation from various tech sources, here are the key points about the anticipated GPT-4.5 and GPT-5 models from OpenAI:
Potential Release Timelines
- GPT-4.5, an incremental update to GPT-4, is rumored to be released around June or summer
5
10
68
Let's forget about OpenAI until GPT-5 drops. The world's gonna change next year anyway. (Seriously, hurry up with the memory update already.)
8
6
69
gpt-5 is soon
("soon" in openai time means 1 soon = a few months)
5
8
69
Reverse engineering the strawberry article from The STaR paper.🍓
3
11
71
GPT-4omni is likely a lightweight, multimodal native model trained on high-quality learning data, with latency prioritized for real-time conversations. Its performance may be better in some areas and worse in others compared to previous models, but it must be tested to experience
5
6
68
"GPT-4.5 or GPT-5 might be an ACI."
Here is a comparison of Artificial Capable Intelligence (ACI) and Artificial General Intelligence (AGI):
Artificial Capable Intelligence (ACI)
- Proposed by Mustafa Suleyman, referring to AI that can pass a modern Turing test by performing
6
16
67
Only OpenAI can give us the dream of AGI/ASI. If they fail, we are finished. We must be their supporters. They are right. I will support them with all my might.
16
6
65
anthropic will likely reach phd level by 2027. google probably too. china won't be left behind either. whatever's gonna happen will happen.
Anthropic CEO Dario Amodei says by 2027 AI models will cost up to $100 billion to train and will be "better than most humans at most things"
92
170
1K
11
3
66
openai has increased lobbying activities with the department of defense since the third quarter of last year, when agi was reportedly achieved internally. the coup by ilya that happened afterward mysteriously ended, and now a former nsa member has joined the board. is the human
8
5
65
"Summary of Gobi and Arrakis by GPT-4o"
Jimmy Apples, a figure known for leaking information about OpenAI's projects, has shared several insights about two AI models, Gobi and Arrakis. Here is a summary of the key points from his leaks:
### Gobi
1. **Sparse Model**: Gobi is
2
11
65
Microsoft and OpenAI are collaborating to build massive supercomputers on Azure to support the development of artificial general intelligence (AGI). This partnership is focused on providing the immense computational resources needed to train the large-scale AI models that could
3
4
61
remembering the exciting rumors about gpt-5 with a refreshed spirit
4
9
62
It's sad to leave OpenAI, but I'm happiest knowing that Ilya is doing well.
After almost a decade, I have made the decision to leave OpenAI. The company’s trajectory has been nothing short of miraculous, and I’m confident that OpenAI will build AGI that is both safe and beneficial under the leadership of
@sama
,
@gdb
,
@miramurati
and now, under the
2K
3K
27K
1
0
63
Claude 3 is superhuman chemists.
Are large language models superhuman chemists?
Large language models (LLMs) have gained widespread interest due to their ability to process human language and perform tasks on which they have not been explicitly trained. This is relevant for the chemical sciences, which
2
23
124
7
10
62
I think either Sam or Ilya was destined to leave OpenAI. No one can deny that without Ilya, the path to AGI would not have been opened, but if the release of GPT-5 were delayed by three years, would you still support Ilya? Even if Sam clung to power, could you blame him for
12
3
61
Even if GPT-5 were to unravel the mysteries of quantum gravity, humans might not be able to verify it.
13
4
61
It appears that OpenAI will effectively lower the price by relaxing the GPT-4 limit, thereby competing with Claude3.😅
6
4
59
Rumor-monger Opus has compiled a list of GPT-4.5 and GPT-5 rumors 🫠:
Based on the recent rumors and speculation, it seems OpenAI may be gearing up to release GPT-5, a major upgrade to its large language model, as soon as this summer in 2024. Here are the key takeaways from the
3
9
56
Since the Safety and Security Committee has been established, it is likely that new models, including GPT-5, will not be announced until the report is compiled in 90 days.
12
5
54
gen-3 alpha catching up to sora in 4 months means that omni might also be caught up by september. anthropic might even release their agentic ai before gpt-5 is released. openai should either speed up its strategy or accept falling to second place.
8
0
55
@elonmusk
@JaengjungG
@OpenAI
if you're really leading the way in open source, how about making grok open source and free to use
3
0
54
please note that sama's "a little patience" spans several years.
thanks a lot for these! some common requests:
AGI (a little patience please)
GPT-5
better voice mode
higher rate limits
better GPTs
better reasoning
control over degree of wokeness/behavior
video
personalization
better browsing
'sign in with openai'
open source
2K
1K
15K
5
1
53
Based on Sam's statements and trademark registrations, there is a possibility that GPT-7 or its subsequent model will be recognized as AGI. Earlier models will likely not be recognized as AGI, regardless of how people feel. In this case, AGI likely refers to something that
9
6
51
AlphaFold 3 by Google DeepMind predicts molecular structures with unprecedented accuracy, improving interaction predictions by at least 50%. It aids in drug discovery and crop resilience, and is accessible globally via the free AlphaFold Server, democratizing scientific research.
Announcing AlphaFold 3: our state-of-the-art AI model for predicting the structure and interactions of all life’s molecules. 🧬
Here’s how we built it with
@IsomorphicLabs
and what it means for biology. 🧵
216
2K
6K
2
2
50
i won't stop promoting claude 3
until openai releases a new model
2
4
48
The smarter the model, the safer it becomes, and AGI will be the safest model.
"Our models have become significantly safer over time. This can be attributed to building smarter models which typically make fewer factual errors and are less likely to output harmful content even
5
8
50
following this formula, we can consider the 2 ooms for algorithm efficiency to be already achieved. if we gain 2-3 ooms for unhobbling with strawberry and agent functionality, then we'll just be waiting for gpu expansion.
people get mad at any model release that’s not immediately agi or a frontier capabilities improvement. think for a second why was this made? how did this research artifact come to be? what is it on the path to?
151
59
1K
1
1
48
We are living in an incredibly primitive era in the course of future history.
staggering
@sama
says GPT4 is "the dumbest model any of you will ever have to use again... by a lot"
full talk:
101
153
2K
6
5
50
Even without GPT-5, we've got our super agent Jimmy Apples 🍎🦸♂️
2
0
48
AGIを商用利用することを認めた初めての発言ではないか？
考察すると、これには重要な意味が含まれています
1. 先行者はその業界を破壊するほどのインパクトと利益を得る
2. 数人で巨大企業を経営できる
3. 大量失業
4. 誰もが高価なAGIにアクセスできるわけではない
1
15
47
if there is agi achieved internally, we should question the timeline of this graph.
6
2
44
patience sam. it will be worth the wait.
people have happily worked so hard to build stuff for you knowing they would never meet you
just hoping that some of the people of the future would continue the quest and build the next branch of the tech tree
1K
1K
15K
1
1
43
OpenAI's rumored Q* model, featuring advanced math and self-improvement abilities, fuels speculation about nearing AGI, amid CEO Sam Altman's unexpected firing and rehiring. (November 24, 2023)
3
5
44
Recent quantum computing progress suggests usable machines soon, with a study showing a 280-qubit processor and improved error correction. This could mean major tech and science advances in ten years.
3
5
44
i understand that everyone wants agi certification soon, but i think this classification is very practical. putting aside the challenge that agi requires system 2 thinking, one of openai's goals is to achieve agi, but it's also something that would replace openai itself. you
10
3
43
Even GPT-5 is kept this much under wraps, it's unlikely they'll ever make AGI widely available. This'll only lead to a power imbalance between those who can access AGI and those who can't. Dystopia is on the horizon. We've gotta tackle this strategically.
3
4
43
Hmm, didn't they stop using numbering? Are there plans to continue with GPT-6, GPT-7 instead of calling it AGI?🤔
13
6
45
Does anyone remember when Elon Musk was arguing that GPT-4 should be open-sourced because it was an early AGI? Now it's called the dumbest model.
4
1
43
Anytime a convo happens around ai with my network in the real world it’s a lot of “ It JuSt MaDe It uP !! “
Then when they show me it’s a green gpt 3.5 logo.
Just retire 3.5 at this point, it’s made a lot of people underestimate future models.
0
41
521
4
2
44
The intelligence explosion can potentially surpass certain computing constraints. However, my optimism remains unchanged :)
II. From AGI to Superintelligence: the Intelligence Explosion - SITUATIONAL AWARENESS ()
3
7
41
Thanks to Google, I was able to sleep for two days after barely being able to sleep at the OpenAI demo.
0
1
40
OpenAI's mission should not be to create AGI, but to bring about the Singularity. They must continue until we can connect with infinite intelligence.
9
3
39
"AI singularity may come in 2027 with artificial 'super intelligence' sooner than we think, says top scientist"
(March 6, 2024)
3
8
38
ok, sam. i trust you completely. chatgpt is 100% safe. you have my guarantee. so go ahead and release gpt-5.
a few quick updates about safety at openai:
as we said last july, we’re committed to allocating at least 20% of the computing resources to safety efforts across the entire company.
our team has been working with the US AI Safety Institute on an agreement where we would provide
785
384
4K
2
0
40
Sam, who raised $7 trillion
Ilya, aiming for super alignment by 2027
roon, who starts talking about myths
Greg, who loves coding
These all indicate that they are on the golden path to AGI/ASI.
3
2
38
i wonder what they have.
2
0
37