AI Leaks and News
@AILeaksAndNews
Followers
683
Following
54
Media
49
Statuses
863
Explore trending content on Musk Viewer
#T20WorldCup
• 1002680 Tweets
#INDvSA
• 484631 Tweets
#TeamIndia
• 355585 Tweets
Bumrah
• 273311 Tweets
South Africa
• 226527 Tweets
#ViratKohli
• 179258 Tweets
Italy
• 166567 Tweets
Switzerland
• 92511 Tweets
#RohitSharma𓃵
• 86491 Tweets
#SvizzeraItalia
• 83207 Tweets
Denmark
• 77548 Tweets
Miller
• 75444 Tweets
#GERDEN
• 58671 Tweets
Alemania
• 57820 Tweets
भारतीय टीम
• 50553 Tweets
ايطاليا
• 49668 Tweets
टीम इंडिया
• 48370 Tweets
Havertz
• 43750 Tweets
Thala
• 43276 Tweets
Spalletti
• 43173 Tweets
Dinamarca
• 41370 Tweets
المانيا
• 40927 Tweets
Musiala
• 38232 Tweets
Suiza
• 36938 Tweets
Sane
• 36918 Tweets
भारतीय क्रिकेट टीम
• 36402 Tweets
سويسرا
• 33843 Tweets
イタリア
• 32467 Tweets
Dortmund
• 30187 Tweets
スロベニア
• 28464 Tweets
#indiawins
• 17246 Tweets
Paul George
• 17147 Tweets
Eriksen
• 15358 Tweets
決勝進出
• 15160 Tweets
Hojlund
• 12872 Tweets
YOUR BANNED
• 12361 Tweets
Danemark
• 10404 Tweets
Pinned Tweet
Wake up
Work
Workout
Code
Learn
Repeat
Do this and you will be more prepared than 99.9% of people for what is going to happen over the next few years
0
1
13
GPT-4 Turbo
Claude 3
Mixtral 8x22b
Gemini 1.5
Udio
Llama3 coming next week
I think we’re accelerating
14
29
250
This GPT4 turbo release seems really rushed/disorganized
No metrics, no concrete examples
What is going on at OpenAI?
28
3
169
If GPT2-chat is just GPT-2 with algorithmic improvements and it performs on par with GPT-4
And each iteration is about a 10x improvement
That means that GPT-5 with these algorithmic improvements would be 1000x better than GPT-4
32
11
146
Sam Altman has confirmed that GPT2-chatbot is not GPT-4.5
13
12
141
ChatGPT is currently down
Could this be the long awaited voice update?
24
6
108
Jimmy expects a second openai release this week
1) End-to-end trained multimodal model, yeah this feels to be Gobi, or in essence Gobi.
Really just confirms that Sam’s following his Iterative deployment plan. We need Agents.
2) Openai didn’t announce everything today. Expect them to likely drop again sometime this week.
0
110
964
10
4
106
ChatGPT is down, with some users reporting momentarily seeing GPT4.5
What does OpenAI have in store for us?
22
6
98
Sam confirms that GPT-5, nor search will be debuted at Monday’s event
Sounds like they’ve postponed their search release
Is it just not good enough to release? Or was it all an elaborate shakedown to figure out who the leakers in the organization are?
not gpt-5, not a search engine, but we’ve been hard at work on some new stuff we think people will love! feels like magic to me.
monday 10am PT.
1K
3K
28K
10
2
91
Google: solves biology
OpenAI: stealth releases im_a_good_boy_69 and im_a_gooder_boy_420
1
4
88
My hope is the gpt2-chat isn’t an OpenAI release, but an unknown lab doing a stealth release to drum up hype
We need new players to apply more pressure to OpenAI
7
2
82
Flowers confirms the search event was a bait and switch to uncover internal leakers at OpenAI
2
9
77
With Ilya turning his attention completely to ASI it is clear that AGI is near and that it is only a matter of time
3
8
77
If the OpenAI demo Monday isn’t for search or GPT-5, what is it for?
Will the officially release the models that have been teased on chatbot arena?
Or will they unveil something we haven’t even heard about?
22
3
72
It’s pretty crazy that OpenAI dropped a model update just to distract us from the fact that they fired two safety researchers for raising safety concerns
10
5
71
OpenAI Paris event next week
Could this be the second release that
@apples_jimmy
mentioned?
We're hosting a developer and founder meet up in Paris next week with the OpenAI team. If you're building with AI and want to attend, drop me a note.
[if you're a tech policy geek that wants to learn about what people are building, drop me a note too]
82
25
309
3
3
65
GPT2-chatbot has returned to the chatbot arena
This time with two versions, im-a-good-gpt2-chatbot and im-also-a-good-gpt2-chatbot
Looks like OpenAI is doing some A/B testing on their models
6
6
64
Rip Google Search
1998-2024
“ One version of the product also uses images alongside written responses to questions, when they’re relevant. If a user asked ChatGPT how to change a doorknob, for instance, the results might include a diagram to illustrate the task, the person said. “
0
33
367
4
6
61
One rumor is that ChatGPT search uses an early version of GPT-5 for novel searches
This makes sense once you consider they’ve already rolled out some version of GPT-5 to enterprise customers
Why wouldn’t OpenAI use their best technology when trying to build new products?
10
6
59
What’s stopping Nvidia from building their own supercomputer to build state of the art AI? They can do it for significantly cheaper than everyone else
19
1
53
Predicting OpenAI releases is essentially impossible at this point. Anyone giving countdowns or dates to potential releases are going to be wrong 99% of the time, and the 1% they are right are completely coincidental. The reasons for this are threefold:
3
4
51
GPT-4o’s elo score on the chatbot arena is trending down
It appears this truly is just a faster GPT-4 class model
I think it’s time for some GPT-5 class models to start rolling out
GPT-4o Elo already tanked below 1300 lol. Knew it was fraudulent...
GPT-5 is the only thing that matters to determine where this stuff is going.
41
17
338
4
1
50
Update: ChatGPT Search has been postponed to next Monday
4
3
49
Microsoft is training a 500B parameter model to compete with OpenAI and google
2
3
47
Some ChatGPT users are getting access to an alpha model that is reportedly significantly better at coding than all other OpenAI models
Holy f*ck it's so good at coding.
It improved on codebase that other models (where gpt4o and gpt2 struggled)
31
54
833
1
4
46
This is the week that the rate of change in the development of AI will truly start to feel exponential
3
10
44
Expecting to see big things on Monday
6
1
41
Which announcement will be bigger on Monday?
Google IO
322
OpenAI’s
777
24
6
42
Elon Musk and xAI are planning a 100k gpu supercomputer
Exclusive: Elon Musk's xAI is planning to build a supercomputer to link 100,000 GPUs to power the next versions of its AI, Grok.
By
@anissagardizy8
11
46
189
0
4
40
Every day that passes makes it more and more likely that OpenAI will just wait until after the American elections to release anything significant
7
0
39
Jimmy apples is expecting agents to be a focus of Monday’s OpenAI event:
1
2
37
Will today’s new projects force OpenAI to release something?
Or will they continue to just show us their new toys that are locked away for enterprise customers only?
4
2
37
What did Dan Schulman see?
Dan Schulman (former PayPal CEO) on the impact of AI
“gpt5 will be a freak out moment”
“80% of the jobs out there will be reduced 80% in scope”
19
38
174
6
1
27
There is no reason for OpenAI to release a search product that is as good as google or perplexity
Monday will be a step change in the capability of search
5
1
31
@futuristflower
So GPT5 is too good, but GPT4 is too bad, so they’re just waiting for everyone to catch up to what they wanted GPT4.5 to be and then they’ll release GPT5?
1
0
32
@futuristflower
There’s a huge disconnect between their incessant preaching of iterative deployment and claiming that the product they’ve provided for over a year sucks and is barely improving but the next thing they have is unbelievable. Their words and their actions do not line up
0
0
25
This is going to be big
5
0
25
I think that it is now clear that OpenAI is significantly ahead of the competition
1
0
25
May is going to be a lot of fun
June is when everything changes
1
2
25
We’re less than a decade away from reaching utopia
I don’t think anyone has truly wrapped their heads around this
5
0
24
Claude 3.5 Sonnet has taken over the top spot on the chatbot arena leaderboard
🔥Breaking News from Chatbot Arena
@AnthropicAI
Claude 3.5 Sonnet has just made a huge leap, securing the
#1
spot in Coding Arena, Hard Prompts Arena, and
#2
in the Overall leaderboard.
New Sonnet has surpassed Opus at 5x the lower cost and competitive with frontier models
29
223
1K
1
1
24
There is a growing amount of evidence that OpenAI is going to release something this week
6
3
23
-OpenAI recently fired several internal leakers who were sourcing all of the information
-OpenAI is actively trying to dispel external leakers by feeding false information
-OpenAI only releases to stymie competitors; they release to counter other’s releases, which dictates timing
1
0
23
Jan Leike has joined Anthropic
Will Ilya follow?
I'm excited to join
@AnthropicAI
to continue the superalignment mission!
My new team will work on scalable oversight, weak-to-strong generalization, and automated alignment research.
If you're interested in joining, my dms are open.
372
532
9K
3
2
22
If Rowan is tweeting about May 9th then it’s a done deal
ChatGPT Search will be here on Thursday
I’d have to imagine it will be the “ChatGPT moment” for search
7
3
22
Some users are already reporting having access to GPT-4o
4
0
20
@futuristflower
I will be disappointed if it is anything less than the “ChatGPT moment” for ai powered search. Anything perplexity level or lower they might as well just post a blog saying “we have it but you can’t use it” just like all of their other tools they’ve made recently
1
0
19
Early indications are that the jump from GPT-4/4o to Claude 3.5 Sonnet is the largest jump in SOTA performance since the the jump from GPT-3.5 to GPT-4
It makes you wonder how powerful Claude 3.5 Opus will be
3
0
20
I think the world would be a better place if we focused more on building more helpful, transformative and safe tools today, rather than worrying about the sci-fi fueled doom scenarios of the tools we make five years from now
4
0
18
Fingers crossed we get an Ilya appearance at the OpenAI event
1
1
17
I don’t see how anyone isn’t team red
Today’s models present little more danger than google search
People need to understand the world is going to change dramatically over the next 5 years and it’s going to be much much worse if that all happens in years 4 and 5
1
1
19
OpenAI is getting back into robotics:
New --
OpenAI is formally resurrecting its robotics team after abandoning efforts to build general purpose robots in 2020.
Its new team has been around for ~2 months and is currently hiring research engineers.
From me and
@kenrickcai
:
15
56
310
2
1
18
Seems to line up with this Sama post
hope you enjoyed!
the new voice mode will be live in the coming weeks for plus users.
we'll have more stuff to share soon :)
365
220
5K
1
2
17
@futuristflower
There’s a dozen ChatGPT clones in the App Store but they’re going after the subreddit?
2
0
16
GPT-4o crushes all other models on the chatbot arena
Breaking news — gpt2-chatbots result is now out!
gpt2-chatbots have just surged to the top, surpassing all the models by a significant gap (~50 Elo). It has become the strongest model ever in the Arena!
With improvement across all boards, especially reasoning & coding
24
224
1K
1
1
16
15 minutes in and so far
#googleio
has just been a demo of the worst text to music model I’ve ever heard
1
1
17
This isn’t what we were looking for…
We're sharing progress toward understanding the neural activity of language models. We improved methods for training sparse autoencoders at scale, disentangling GPT-4’s internal representations into 16 million features—which often appear to correspond to understandable concepts.
344
861
5K
0
2
17
Jan Leike speaks on why he left OpenAI:
0
1
16
We’ve been talking about Q* since November, yet nobody has released a model utilizing it. Why is that?
5
0
16
Claude 3.5 Sonnet surpasses GPT-4o
Introducing Claude 3.5 Sonnet—our most intelligent model yet.
This is the first release in our 3.5 model family.
Sonnet now outperforms competitor models on key evaluations, at twice the speed of Claude 3 Opus and one-fifth the cost.
Try it for free:
431
2K
7K
0
0
15
This in no way is to disparage leakers.
@apples_jimmy
@futuristflower
and all other true leakers are incredible and we truly aren’t thankful enough for what they do. It’s just that their job as leakers have been made incredibly more difficult due to the current circumstances
2
0
16
Nothing is set in stone and these companies love to shake things up at the last minute to expose leakers but it’s currently looking like
-Gemini update tomorrow
-ChatGPT voice release Thursday
Today’s outages also might have some effect on the releases as well
0
1
16
Very soon OpenAI are going to have artificial intelligence that is capable of performing novel science.
When this happens are they going to distribute this science? Will they publish papers? Will they partner with other scientific groups to publish their research?
2
2
15
Anthropic teases Claude Sonnet 3.5
Fc zbvx ts temxnsq nx mzog jlbuv gusn zofg hhfs:
Ebwxk vnii mzceaw tfr fvpowf sbyglovaw fmr,
Nyp fgrryw xjf, lrx 'mosra xaw huvvw sbq ssnjhu'f vnf,
Che rxo qeremaca ophgaf, n abkse oyw.
KEY:
95
79
820
1
0
14
Not only are we all going to make it, but it’s going to be more beautiful than you can imagine
0
0
14
@futuristflower
That position made sense up until Microsoft started building a $100B computer, I can’t imagine they make that bet unless they were certain it will yield AI beyond AGI
1
0
13
That definitely did not disappoint
Google better come out with something big tomorrow
1
2
13
We are accelerating*
Google presents Leave No Context Behind: Efficient Infinite Context Transformers with Infini-attention
1B model that was fine-tuned on up to 5K sequence length passkey instances solves the 1M length problem
28
260
1K
0
1
13
The question is how is SSI going to get the compute necessary for ASI?
7
0
13
Ilya Sutskever unveils his new company
Superintelligence is within reach.
Building safe superintelligence (SSI) is the most important technical problem of our time.
We've started the world’s first straight-shot SSI lab, with one goal and one product: a safe superintelligence.
It’s called Safe Superintelligence
1K
2K
14K
0
2
13
WWDC predictions:
-more agentic Siri than chatbot Siri
-only available on new models
-won’t actually release until fall/winter
1
1
12
Official Ilya update:
Ilya and OpenAI are going to part ways. This is very sad to me; Ilya is easily one of the greatest minds of our generation, a guiding light of our field, and a dear friend. His brilliance and vision are well known; his warmth and compassion are less well known but no less
2K
2K
26K
0
1
12
Can someone from OpenAI actually explain what GPT2-chatbot is?
Free marketing from a stealth release is cool and all, but people need to understand what the models we are using are
We need more openness from “Open”AI
0
0
11
@futuristflower
@OpenAI
At this point I think it would take more than that. I’ll use voice, but 3.5 Sonnet is a better text model than anything they have out.
1
0
11
Following Nancy Pelosi’s stock trading might be the best alpha on upcoming releases in AI
BREAKING 🚨:
Nancy Pelosi just bought $5M of the AI company Databricks
Unfortunately, Databricks is a privately held company and not available to be bought by the public
Sorry people, you don’t have access to this one.
291
2K
15K
1
0
10
Very interested to see what Google has for us tomorrow
0
1
11
I’m surprised that OpenAI released and upgrade to GPT4 turbo rather than putting out a Haiku competitor
Especially with two small Llama3 models coming next week
It looks like OpenAI is content with falling to third place in the small/cheap model race
2
0
10
30 minutes to go
Here is the link to WWDC
All eyes on Apple at today’s WWDC! 📱🍏
With the tech world buzzing about AI, the pressure is on for Apple to deliver some groundbreaking announcements.
Catch it live at 10 AM Pacific Time:
0
4
10
1
2
9
Here is where you can access the mystery gpt2 chat model:
GPT-4.5 probably, 8 prompt limit. Have fun. It's good.
model: gpt2-chatbot
71
134
1K
0
0
10
We currently aren’t close to AGI in terms of capabilities, but the path is exponential
When we are remotely close to AGI in terms of capabilities, we will be extremely close in terms of time
The same is true of ASI
0
1
10
Meta is planning to not open source Llama3 400B
Meta plans to not open the weights for its 400B model.
The hope is that we would quietly not notice / let it slide.
Don’t let it slide.
1
152
1K
1
1
10
@futuristflower
Possible outcomes:
Push the button:
50% chance everyone dies
50% chance we solve aging and disease
Don’t press the button:
100% chance everyone eventually dies
5
0
10