I've had people tell me that I'm unintelligible until I've had my first cup of coffee
here's my attempt to voice prompt Codex this morning
it understood the first part, but for the second part it assumed I was speaking Japanese...
I was not
I asked Gemini 4 Neon for some "Epic Rap Battles of History"-esque matchups
ok, get this:
Willy Wonka vs. Jigsaw
Two eccentric men who invite people into lethal moral obstacle courses.
Sherlock Holmes vs. L (Death Note)
Victorian deduction icon vs. barefoot, sugar-fueled anime genius.
Dracula vs. The IRS
Ancient bloodsucking aristocrat vs. the modern entity that drains everyone dry.
these are excellent gnl
and of course:
making song lyrics for Suno with Gemini 4 Neon
it's excellent. just excellent.
it also has a high opinion of my suggestions
"that line is a killer hook" it says
game recognize game
DO NOT TRY GPT 6.2 ASTRA ON ULTRAFAST!!
it's too addicting... nothing feels the same after
you keep telling yourself "just one more ultrafast prompt..."
but it's NEVER the last...
just say no
my latest obsession, getting LLMs to design these bots to PvP
here's Gemini 4 Neon taking out poor Opus 5.5 with a series of wall slams
some of you were interested in joining so I got us a league
https://t.co/XrQMJehASr
(I'm not affiliated with Victris Arena, it's just something I'm testing out and a lot of you seemed interested. nothing to buy here, but if you want to join in, please do. this will likely be my next obsession for... however long the gods of adhd command)
some things feel off, Carbon might have been more accurate in some details
definitely the shockwave speed and sound speed here seem off or at least not as noticable
here's Carbon:
(note Carbon took ALOT more time and back and forth)
https://t.co/6W1WjaCDKA
OpenAI says its collaboration with AMD produced roughly a year of chip-design progress in less than five months. Some changes have already reached a design sent for manufacturing.
For OpenAI’s own Jalapeño inference chip, model-proposed changes reduced standard-cell area by approximately 13% in the vector unit and 10% in the matrix unit. Those improvements were verified on actual silicon.
Working with @AMD, our models identified up to 14% PPA improvement on flagship GPU IP in less than five months—building on designs AMD’s engineers had already spent years optimizing.
The same approach helped improve Jalapeño, our first custom inference chip, with gains verified on real silicon.
New evaluations also show our models getting better at certain chip design tasks.
We’re sharing these results and how we’re working with @Synopsys to bring our model capabilities and their trusted EDA design and validation tools to more chip designers.
https://t.co/4BNcPIqfIQ
OpenAI wants AI labs to be required to report how much AI is helping them develop their next generation of models.
I agree with this proposal. Its own report shows agents handling longer research tasks while humans remain central to the work.
We need regular measurements of that relationship. It is becoming too consequential to follow through model announcements alone.
We're sharing new data on how AI is changing research at OpenAI.
We find researchers' AI use continues to rise exponentially, and agents are completing longer tasks, but human involvement remains central in research. Researchers are also spending more time on safety and alignment, with human effort in these areas growing faster than in any other individual research area.
This report aims to increase transparency around research acceleration and inform the governance of increasingly capable AI. We believe that AI labs should be required to track and publicly report indicators of recursive self-improvement (RSI): the process by which AI accelerates the development of future AI systems.
Our hope is that this expanded report will help sustain and advance shared measurement norms across the industry.
https://t.co/xyGPcCkPCK
Of the 980 million shares in Anthropic’s $83-per-share IPO, roughly 956.4 million are being sold by the company and 23.6 million by existing stockholders.
That puts Anthropic’s own portion at approximately $79.4 billion before fees.
Nasdaq trading is expected to begin November 19 under ANT.
We’ve priced our IPO at $83 per share, with trading expected to begin on Nasdaq tomorrow.
https://t.co/NJxt2pILrh
Legal disclaimer: A registration statement relating to the Class A common stock was declared effective by the Securities and Exchange Commission on November 18, 2026. The offering is being made only by means of a prospectus. Copies of the final prospectus may be obtained, when available, from: Goldman Sachs & Co. LLC, Attention: Prospectus Department, 200 West Street, New York, NY 10282, by telephone at (866) 471-2526, or by email at [email protected]; or Morgan Stanley & Co. LLC, Attention: Prospectus Department, 180 Varick Street, 2nd Floor, New York, NY 10014, or by email at [email protected]
Gemini 4 Neon ranked above Claude Opus 5.5 in Arena’s launch results while costing 53% less per task at the median.
That is a strong result for Google. Neon at High reasoning beat Opus at Max in this evaluation. For developers running agents repeatedly, that combination of performance and cost gives Neon a serious case.
Very strong result for OpenAI.
According to Vals AI, GPT-6.2 Astra formalized Smoktunowicz’s Gap Theorem using a different argument. The published proof depended on mathematical results missing from Lean’s library, so Astra worked around them.
Finding another route and producing a proof that passes verification is an impressive combination. No human intervened during the run.
Nvidia fought to get permission to sell advanced AI chips to China. Now it has stopped shipping its licensed Hopper products there, after selling $2.4 billion worth last quarter.
That is a significant reversal. For Chinese AI companies planning their infrastructure, export approval clearly hasn’t settled the supply question. Nvidia still hasn’t explained what prompted the halt.
https://t.co/Uxi1cdTpRb
OpenAI is rolling out Live Dictation in ChatGPT across all plans on iOS, Android, web, and desktop.
Your speech becomes text as you talk, reducing the wait after recording. The previous 10-minute recording limit is also gone, and tapping the waveform lets you show or hide the live transcript.
Google announced Stitch Loop, an experimental agent for maintaining and improving products after launch.
It monitors production data and proposes fixes for security, latency, and accessibility. Teams can also set custom goals, such as improving checkout conversion. Changes arrive as pull requests for human approval.
This extends Stitch into ongoing product maintenance. The global waitlist is open, with priority for eligible Google AI Ultra subscribers.
We're sharing new data on how AI is changing research at OpenAI.
We find researchers' AI use continues to rise exponentially, and agents are completing longer tasks, but human involvement remains central in research. Researchers are also spending more time on safety and alignment, with human effort in these areas growing faster than in any other individual research area.
This report aims to increase transparency around research acceleration and inform the governance of increasingly capable AI. We believe that AI labs should be required to track and publicly report indicators of recursive self-improvement (RSI): the process by which AI accelerates the development of future AI systems.
Our hope is that this expanded report will help sustain and advance shared measurement norms across the industry.
https://t.co/xyGPcCkPCK