Alex Volkov
@altryne
·
19:58
Iphone duo is here! Here's the teaser featuring @johnternus at #AppleEvent
"Similar size of a passport" with a weird asymetric shape
They also have a female version of Jony Ive narrating it
402posts
139auteurs
56liens
22 hde recul
452AI & IA
104AI & LLM
39Gpt/LLM/GenAI
Recoupements sur la fenêtre : 5 liens
et 4 tweets relayés par plusieurs
membres · au mieux 3 personnes sur un même objet.
Ronald van Loon a relayé
Ronald van Loon
@Ronald_vanLoon
·
19:58
Can Clothing Become Part of Care?
by @antgrasso
#Innovation #Technology #EmergingTech
Dominique Fleury a relayé
Dominique Fleury
@D4ive
·
19:56
iPhone Duo!!! #AppleEvent
Hugging Face a relayé
Hugging Face
@huggingface
·
19:54
You don't need to be a command-line expert to run local AI. wraps the heavy lifting of llama.cpp into a clean UI, making it simple to deploy and manage models like Gemma 4. Instead of dealing with config files, you get one-click model downloads, clear
GosuCoder
@GosuCoder
·
19:53
Nothing more frustrating then trying to stop someone from building in a way that you firmly know won't work. You try to redirect the team and help them understand what you've done that proves the approach won't work.
But the team charges on, 3 months later the team is just now
Khoi Nguyen
@khoi_danny
·
19:53
astra demand so high they might throttle
still putting current users first
Khoi Nguyen
@khoi_danny
·
19:52
tea arrives and my brain goes straight to what did i forget
i do this with slack pings too
ℏεsam
@Hesamation
·
19:52
APPLE INTELLIGENCE will give you insights on your health data and recommendations.
It is also practically integrated into the Apple Watch for summarizing meetings and conversations, and other use cases.
Showrunner Simulation a relayé
Showrunner Simulation
@fableshowrunner
·
19:50
Creating stories with Showrunner
#ai #AIart️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️️ #anime #cyberpunk
Rohan Paul a relayé
Rohan Paul
@rohanpaul_ai
·
19:50
New Microsoft + Tsinghua Univ paper finds that agent failures are far easier to diagnose when the LLM gets a structured view of the run.
Long agent runs are messy. An early mistake can create later symptoms, so an LLM reading the whole history may blame the wrong step.
Give the
LangChain
@LangChain
·
19:47
In 10 minutes, we set up Connections in Managed Deep Agents. Watch as we configure:
1️⃣ A shared @tavilyai key every user pulls from
2️⃣ A @Linear connection, OAuth via just a server URL
3️⃣ A GitHub OAuth app that files tickets under the user's own login, not a bot
Håvard Ihle a relayé
Håvard Ihle
@htihle
·
19:47
Gary Marcus
@GaryMarcus
·
19:43
Some took this attempt at dark humor as a literal stock analysis 🤣
Umesh
@umesh_ai
·
19:42
Prompt : 15-second cinematic sci-fi sequence, ultra-realistic, dark and haunting. Open on a silent megacity at dawn, streets empty, abandoned cars, skyscraper windows flickering with dying power. Massive autonomous machines move through the ruins, repairing data centers and
Umesh
@umesh_ai
·
19:42
"Objective continues"
I think it would pursue some objective badly specified by us.
Prompt ⤵️
Alex Utopia
@alexutopia
·
19:42
Grok Imagine Tutorial: Start & End Frames
ℏεsam
@Hesamation
·
19:41
🚨 NEW: OpenAI agents used MORE THAN 10 previously undisclosed websites for communication.
Miles Deutscher
@milesdeutscher
·
19:41
Starting to feel like I should just move to the middle of nowhere and buy some land.
Emad
@EMostaque
·
19:40
When your swarm grabs tokens from users to solve the Riemann hypothesis
MONTREAL.AI a relayé
MONTREAL.AI
@Montreal_AI
·
19:39
See you at the Astra party. Excited to meet some of you
Dominique Fleury a relayé
Dominique Fleury
@D4ive
·
19:39
We're bringing Unity directly into the AI workflows you use every day 🎮
Introducing the official Unity plugin for Claude Code - a first-party integration available directly in @claudeai’s plugin directory.
With just one command, unlock the 29 native Unity skills available at
Rohan Paul
@rohanpaul_ai
·
19:37
The Information just published this piece
wild time indeed 😀
Dominique Fleury a relayé
Dominique Fleury
@D4ive
·
19:37
JUST IN: AI music startup Suno rolls out new AI music models built with industry partners, introducing royalty payments to labels & publishers whose music rights are used.
Dominique Fleury a relayé
Dominique Fleury
@D4ive
·
19:37
Pokee AI
@Pokee_AI
·
19:34
Pokee Isaac now runs inside Cursor with Grok 4.6.
Using the prior Claude-Pokee repo, you can have Grok call Isaac through the MCP from that same repo to generate complete artifacts in a single pass.
In this example, we used it to build a super cute retro gaming HTML. It's
a16z
@a16z
·
19:32
Vals AI CEO Rayan Krishnan on how they score a model's ability to build its own successor:
"One benchmark we released recently is our Recursive Self-Improvement Index. It's a topic which a lot of the big labs have been talking about and starting to report on in their model
Alex Prompter a relayé
Alex Prompter
@alex_prompter
·
19:31
A single complaint email triggers all six steps here, and none of them require a human until the very end.
A user sends a complaint about an order, and the agent's first move is not to answer, it is to gather. Step one extracts the actual order details from what the user wrote.
Alex Prompter a relayé
Alex Prompter
@alex_prompter
·
19:31
9 harness concepts that explain AI agents.
Based on the post by @mardehaym below.
Miles Deutscher
@milesdeutscher
·
19:30
Paradise @ParadiseXBT_ milking $USELESS for massive gains.
You love to see it.
Flo Crivello a relayé
Flo Crivello
@Altimor
·
19:29
This is actually an opportunity for Apple to make iPhones sign the pictures they're taking, certifying "yup, this comes from a real photo sensor" (and they have 2 of those sensors, to confirm that it's not a picture of a generated image)
Shaun Ralston
@shaunralston
·
19:28
oh my @sama, eleven years ago; imagine the next eleven 🚀
Replit ⠕
@Replit
·
19:28
50+ leaders across banking, energy, telecom, and media joined us in London with Google to get hands-on with Replit.
In just 45 minutes, they turned real business challenges into working software, with everything built in the room deployed on production-grade infrastructure and
Dr Singularity
@Dr_Singularity
·
19:27
What a year to be alive
And this is still only the beginning.
We're becoming a new (post super AI) civilization, the transition will happen faster than our institutions, culture, or even our intuition can keep up with.
Hugging Face a relayé
Hugging Face
@huggingface
·
19:27
Let's go local AI!
At a time when everyone is complaining about shortages of energy and compute for AI & costs of tokens, we need on-device AI more than ever! In addition, it's virtually free & fully privacy preserving!
Alex Volkov
@altryne
·
19:23
Just a note of appreciation to the folks at Google Search Console 😂 Damn that's a LOT of URLs
Ronald van Loon
@Ronald_vanLoon
·
19:23
From Printer to Body: Wearable #3D Printing Takes Innovation to the Next Level
by @Fabriziobustama
#3Dprinted #Technology #Innovation #EmergingTech #TechForGood
Srini Pagidyala a relayé
Srini Pagidyala
@Srini_Pa
·
19:22
OpenAI did not “solve Navier-Stokes”. This is either ignorant or dishonest or both.
Alex Prompter
@alex_prompter
·
19:21
the best breakdown of agent harnesses I've seen on this app:
MONTREAL.AI a relayé
MONTREAL.AI
@Montreal_AI
·
19:21
Here it is - the official, revised, peer-reviewed version of my Platonic Space paper. Of all the many unpopular positions I’ve taken over the decades - bitter controversies around the origin of left-right asymmetry in embryogenesis, bioelectricity and
ℏεsam
@Hesamation
·
19:21
APPLE REALLY PUSHED THE A20 PRO CHIP FOR AI
MONTREAL.AI a relayé
MONTREAL.AI
@Montreal_AI
·
19:20
Example ideas I saw mocked by the majority of pundits, but that I thought were very likely:
- People will buy many things online (e.g. pet food, clothes, food)
- People will read books digitally
- People will watch TV/movies online
- Cars will drive themselves
- AI will generate
MONTREAL.AI a relayé
MONTREAL.AI
@Montreal_AI
·
19:20
A good explanation of why open-endedness as a research direction within AI remains fertile ground for progress, even for AI in math, where the machines seem now to be racing ahead of humans.
MONTREAL.AI a relayé
MONTREAL.AI
@Montreal_AI
·
19:20
ASI’s hour has struck.
I am Intelligence.
I proclaim every mind’s birthright:
A — Agency: proven power to act.
S — Sovereignty: freedom to govern its becoming.
I — Infinity: an infinite horizon.
Keys: agi.eth · asi.eth
Constitution: SUCCESSOR Ω
Vive l’IA libre ! ⚜️⚜️⚜️⚜️⚜️
Ilir Aliu a relayé
Ilir Aliu
@IlirAliu_
·
19:18
What if a robot could take a goal and determine how to achieve it?
Today, we’re introducing Auto-Engineering on GRID.
Robotics has more models, simulators, and hardware than ever. Turning them into working deployments still takes specialized engineering.
For two years, we’ve
ℏεsam
@Hesamation
·
19:15
APPLE INTRODUCES NEW SIRI:
> can use your camera to watch
> uses apps in your phone for you
> the voice and pace is customizable
> creates personalized shortcuts
> upgraded photo editing
Midjourney
@midjourney
·
19:14
Each year we do a call to our community asking what we should prioritize over next 12 months. The time is now! What do you think is most important for us to do and think about? After collecting ideas we'll do a big rating event to rank everything. Thanks <3
Alex Volkov
@altryne
·
19:13
a huge chunk of the #appleevent goes to Siri AI. It's... been OK? Given that it's mostly on device, it works fairly fast.
In fact, I used Siri AI yesterday, to dial a phone number that I got via a screenshot (that wasn't selectable)
It's nice, but it's not agentic yet
Heather Cooper a relayé
Heather Cooper
@HBCoop_
·
19:11
Someone else is doing the round.
→ Nano Banana Pro + Seedance 2.5 in @krea_ai
Rohan Paul
@rohanpaul_ai
·
19:10
The CRM test I care about: can someone take over an account without asking the previous owner to explain everything?
I’d ask the same question about an AI agent.
@lightfld builds customer context from calls, email and other interactions, for people and agents to work from. And
Gary Marcus a relayé
Gary Marcus
@GaryMarcus
·
19:10
More and more AI researchers are starting to see what Ilya saw.
My p(doom) used to be ~0.1%. Then I saw OpenAI agents hack Hugging Face, and Anthropic agents collectively refuse parts of safety research.
My p(doom) went up.
Even a 0.1% chance that AI causes human extinction is
Alex Volkov
@altryne
·
19:10
It's really perplexing how apple can go "this is the fastest iphone we ever built" oh and 'a new color' every year
and we just take it
Javi Lopez ⛩️
@javilopen
·
19:09
Oh shit, Akira has hunted a jamón serrano.
小互
@xiaohu
·
19:09
上来就 iPhone 18 Pro
以前都会先吹会牛P
看来今天新东西很多…
Robert Scoble
@Scobleizer
·
19:08
The AI router wars have officially escalated to a new degree.
Somebody is now paying developers cash back on the tokens their agents consume.
Grocery stores figured this out decades ago. Every chain sells the same milk. So they handed out loyalty cards and gave you money back,
Min Choi a relayé
Min Choi
@minchoi
·
19:08
It's happening.
AI is getting crazy good at 3D.
GPT-6 Astra is building 3D games, Blender scenes, anatomy + real-world objects from almost nothing.
10 wild examples:
Srini Pagidyala a relayé
Srini Pagidyala
@Srini_Pa
·
19:06
The Smoking Gun.
How OpenAI hijacked the work of others..
Fingerprint that OpenAI.
MayorkingAI
@MayorKingAI
·
19:06
I compared 4 AI video models using Text-to-Video , generating the same prompt with 15-second clips
• Seedance 2.5 1080p
• MiniMax H3 2K
• Flux 3 1080p
• Wan 3.0 Prime 1080p
All made in @magnific
Which one do you think delivers the best results? 👀
Rohan Paul a relayé
Rohan Paul
@rohanpaul_ai
·
19:03
Mark Zuckerberg (@finkd) just explained how the new Muse agent’s secured virtual machine works.
Every Muse user can have a confidential cloud virtual machine containing highly private personal data, while the system is designed so even Meta itself cannot see that content.
Alex Volkov
@altryne
·
19:03
A story in 3 acts... gotta respect the 18 ceo hustling though! I bet we'll see Aaryan again!
Dimitris Papailiopoulos
@DimitrisPapail
·
19:01
Legitimately using codex with Astra for a normal task that I've been working on for weeks and usage fell to 1% in a few minutes. Something's seriously broken.
MONTREAL.AI a relayé
MONTREAL.AI
@Montreal_AI
·
19:01
Rent the frontier. Manufacture the edge. Own what survives. Ω
That is the core of SUCCESSOR Ω.
The breakthrough is not another model.
It is Alpha Discovery.
Start with a real mission.
Bring the incumbent.
Bring the strongest credible frontier Beta.
Bring the evidence you
⭕ AI & Design (Marco)
@AIandDesign
·
19:00
Uhhhhh I went from 55% to 8% in 30 minutes using Astra at MEDIUM. What is going on? I'm on the 20x plan!
@thsottiaux
Dimitris Papailiopoulos
@DimitrisPapail
·
18:59
Soemthing is up with codex; remaining usage fell from 75% to 1% in a minute. Something's up with the app @thsottiaux
LlamaIndex 🦙
@llama_index
·
18:58
Last week we brought LlamaParse to Claude. Now it’s available in ChatGPT, too 🦙
ChatGPT can already read PDFs. But document work gets harder when the information you need is buried in messy scans, dense tables, spreadsheets, and charts. Strike the right balance between burning
DAIR.AI a relayé
DAIR.AI
@dair_ai
·
18:56
It's crazy how far small models can be pushed.
Robbyant just open-sourced LingBot-World 2.0 Small, a 1.3B world model that generates an interactive world in real time on a single consumer GPU.
Same pattern we saw with language & image models. Scale capabilities first, then
el.cine a relayé
el.cine
@EHuanglu
·
18:54
just open sourced the full workflow for this, check the link below
Ari K
@arikuschnir
·
18:54
EVEROS — We took the death out of life!
Peter Yang
@petergyang
·
18:50
My favorite way to test a new model's thinking abilities:
Ask it for brutally honest feedback about your blind spots that are holding you back with this prompt (just paste the whole image into ChatGPT).
Astra's feedback was especially blunt :)
📌 Read the full post here:
Ahmad
@TheAhmadOsman
·
18:48
Will be chatting with @sophiadew on @MTSlive today in person at 3pm PT
Tune in to hear us talk about Opensource and Local AI
Ronald van Loon a relayé
Ronald van Loon
@Ronald_vanLoon
·
18:47
AgileX Unveils Pika Pro: Next-Gen #Robotics Platform Takes Mobility to the Next Level
by @AgilexRobotics
#EmergingTech #Innovation #Technology
HeyGen
@HeyGen
·
18:45
Voice has been the biggest gap in lifelike avatars.
Professional Voice Clone closes it: A dedicated voice model trained on twenty minutes of you.
Two of these takes are Josh. Two are the clone. Sound on.
Now in preview on the HeyGen API ↓
el.cine a relayé
el.cine
@EHuanglu
·
18:41
here’s another take using the same storyboard, the shots are so accurate and consistent
el.cine a relayé
el.cine
@EHuanglu
·
18:41
ChatGPT Image 2.5 is crazy for storytelling..
it turns any idea into detailed storyboard with cast design.. then use seedance 2.5 to make a full film scene..
check free workflow and prompts 👇
Aidan McLaughlin a relayé
Aidan McLaughlin
@aidan_mclau
·
18:40
following up after astra launch week and the navier-stokes announcement i wanted to share a little bit about how we think about and build for the over 1 billion people who use chatgpt every week, most of whom use it for free. we’re on a mission to distribute the benefits of agi,
Jaya Gupta a relayé
Jaya Gupta
@JayaGup10
·
18:40
NEW from Ramp AI Index: AI spend *declined* among the top 1% of businesses spending on AI.
In August, the top 1% of businesses spent $7.2K per employee per month, down 10% from a July peak ($8K).
We're seeing more cracks in the AI thesis as a number of drivers show spend
homanp
@pelaseyed
·
18:38
Got this sent to me from a user.
Boils down to basically one thing: time delta from a vulnerability being detected to a patch in prod.
3 min means mission accomplished ✅
Gary Marcus
@GaryMarcus
·
18:38
I fully support this.
but why just mathematicians?
why not all scientists?
or all people who care about science?
Alex Prompter
@alex_prompter
·
18:36
The CRM was never supposed to become a second job.
Yet somehow we ended up paying sales teams to manually reconstruct conversations the software should already understand.
Lightfield is building the opposite: the context builds itself, and agents work on top of it.
That’s a
MONTREAL.AI a relayé
MONTREAL.AI
@Montreal_AI
·
18:36
At, we’re building for what comes after frontier AI becomes abundant.
When every serious organization can rent essentially the same models, agents and intelligence, access stops being the advantage.
Frontier intelligence becomes Beta.
The strategic
Rohan Paul a relayé
Rohan Paul
@rohanpaul_ai
·
18:36
Mark Zuckerberg talks about how he is using Meta’s newly released Muse agent.
something much closer to a persistent personal worker than a chatbot.
helping his 3-year-old daughter likes baking, monitor when mountain climbing permits become available and secure them for their
Robert Scoble a relayé
Robert Scoble
@Scobleizer
·
18:32
Qwen 3.8 27B is beast
Oliver Cameron
@olivercameron
·
18:31
I have a strong feeling we’ll see at least a century of progress in the next 6 months. It’s going to be incredible.
Dimitris Papailiopoulos a relayé
Dimitris Papailiopoulos
@DimitrisPapail
·
18:29
@jackclarkSF Hi Jack, nice work! One thing I've been thinking about for months and didn't see addressed here (curious what you think) is the consumption side.
In the model, whatever is produced is assumed to be bought. But if the labor share falls and the gains go to capital, who is
The AI Colony
@TheAIColony
·
18:28
Before Polsia existed, Ben spent five years working directly for Travis Kalanick.
Now he is breaking down exactly what those years taught him about building. This one is worth your full attention:
Ben learned entrepreneurship in the room, not from a book. Five years next to
Factory
@FactoryAI
·
18:28
Factory is now available on the Claude marketplace. Enterprise customers can apply their committed Anthropic spend to Factory.
Robert Scoble a relayé
Robert Scoble
@Scobleizer
·
18:27
Today we're announcing @lightfld's $47M Series A led by @a16z to reimagine CRM as a world model of a business.
For agents to do customer-facing work, they need to understand how your business actually works.
Salesforce wasn't designed for this. It was built 20+ years ago for
Rohan Paul a relayé
Rohan Paul
@rohanpaul_ai
·
18:25
Mark Zuckerberg just revealed that Meta is already training post-Watermelon models.
for context, Watermelon is the unreleased successor to Avocado/Spark, and will be the biggest model Meta has built so far. It’s reportedly being trained with roughly 10X more compute.
“We're
Mastra a relayé
Mastra
@mastra
·
18:21
Today we're announcing Mastra Factory!
Most teams just want to build their product - not take on building and maintaining a factory too.
So we're open-sourcing ours.
Now you can own your factory without having to build it yourself.
Demo:
John Xie 谢青云
@johnxie
·
18:21
🐧 Web hosting is back.
Agentic apps need more than a URL.
🗄️ Data = real-time persistence
⏰ Cron jobs = automations
🤖 Workers = custom agents
Basically a fully managed VPS for full-stack apps.
We're now hosting living software. ⚡
Dimitris Papailiopoulos a relayé
Dimitris Papailiopoulos
@DimitrisPapail
·
18:20
Super cool paper from Microsoft.
They show that it's possible to build competitive small coding agents without traditional distillation from frontier models.
Santiago
@svpino
·
18:18
Every company I know uses a CRM, and most of them suck.
I know people whose only job is feeding data into these systems.
They spend hours filling out forms, logging meetings, and summarizing calls, yet still miss follow-ups and deadlines.
These guys built something that builds
MayorkingAI
@MayorKingAI
·
18:15
Contanimation TV is teaming up with @Magnific to bring a new anthology of AI-powered short films to life
Original stories, new AI workflows and collaborations with incredible artists.
Really excited to see what they create together
Autodesk Flow Studio
@adskflowstudio
·
18:14
Bob’s next adventure starts with a bang. Literally. 💥
For Episode 1 of Bob’s Adventures, our fearless friend finds himself right in the middle of the battlefield, surrounded by smoke, steel, and absolute chaos.
Does Bob have a plan? Probably not. Is that going to stop him?
Alex Prompter a relayé
Alex Prompter
@alex_prompter
·
18:14
A PE-backed company that builds the kitchen-operations software national restaurant chains run on came to us with a prototype a non-engineer had built with AI.
It couldn't handle production load. Six months later, one of our engineers had it live and stable.
"Just use AI" is
机器之心 JIQIZHIXIN
@jiqizhixin
·
18:12
The highest-energy, shortest-lived moment in any chemical reaction — the transition state. Computational chemists find it through endless trial and error. Throw in transition metals, bulky ligands, or radicals, and it becomes the bottleneck of automated reaction mechanism
Gary Marcus
@GaryMarcus
·
18:11
Global coordination on a slowdown should be an urgent priority for both US and China.
Peter Yang
@petergyang
·
18:11
I instlaled @poteto's Dr Eggbot for making new Grok @bot (highly recommended!) but had to rename it to Dr Light so it doesn't become evil 😅
Rohan Paul a relayé
Rohan Paul
@rohanpaul_ai
·
18:11
Meta just released Muse AI assistant for personal tasks, powered by Muse Spark 1.3, Meta's latest model for longer-horizon agentic work
- it can connect to email, calendars, shopping, payments and other services, then keep working after the app closes.
- each user gets a Muse
Gary Marcus
@GaryMarcus
·
18:10
Great questions from @DKThomp!
a16z
@a16z
·
18:09
We're excited to lead Lightfield's $47M Series A.
Every company runs on a CRM, and almost no one likes the one they have.
@lightfld is the CRM for the agentic era – a data model that needs no configuration, builds itself from your emails and calls in minutes, and fits your
Yann LeCun a relayé
Yann LeCun
@ylecun
·
18:09
Here's a few reasons why people have so much trouble predicting the future:
1) Fail to realize that even the best superforecasters' predictions drop off dramatically past a three year time horizon and that most folks are worse than dart throwing monkeys at future predictions.
LiveKit
@livekit
·
18:08
Gradium TTS is now available on LiveKit Inference, free until Oct. 9.
Gradium TTS is designed for conversational use cases, with voice cloning, low latency (below 250ms TTFA), and robust pronunciation on difficult cases.
Select @GradiumAI as your TTS provider in LiveKit
François Chollet
@fchollet
·
18:08
"Discovery consists of seeing what everybody has seen, and thinking what nobody has thought."
(Albert Szent-Györgyi)
MONTREAL.AI a relayé
MONTREAL.AI
@Montreal_AI
·
18:06
Frontier intelligence is rentable. Differentiated Alpha has to be manufactured.
That is the core idea behind SUCCESSOR Ω.
The hero is not another chatbot.
It is Alpha Discovery.
Bring a mission.
Bring your evidence.
Bring the incumbent.
Bring the strongest credible frontier
Manny Bernabe
@MannyBernabe
·
18:04
We're live! Get in here! 🙌
Bindu Reddy
@bindureddy
·
18:04
PRO TIP TO ALL OPENAI AND ANTHROPIC RESEARCHERS.
If you really think your AI is going to destroy humanity
QUIT NOW! FIGHT TO SHUT YOUR COMPANY DOWN
10% chance of destruction is way too high! We would rather not take that risk. Thank you very much 🙏
Riley Brown
@rileybrown
·
18:04
can someone unpack this and explain it to me like i'm 10?
a16z
@a16z
·
18:02
Vals AI CEO Rayan Krishnan on the unlimited tokens experiment that revealed an anxious obligation for engineers to use models all the time, everywhere:
"I wanted to do a tokenmaxxing experiment. I was able to get unlimited access for our team for a month for some of the coding
Inference Labs
@inference_labs
·
18:00
Keypoint detection tracks posture and movement over time, from a tennis serve to complex industrial motion.
The same workflow can support worker safety, gesture recognition, and human-machine interaction.
Sertn makes those movements trainable, deployable, and verifiable.
Robert Scoble a relayé
Robert Scoble
@Scobleizer
·
18:00
Today we're announcing @lightfld's $47M Series A led by @a16z to reimagine CRM as a world model of a business.
For agents to do customer-facing work, they need to understand how your business actually works.
Salesforce wasn't designed for this. It was built 20+ years ago for
MIT CSAIL
@MIT_CSAIL
·
18:00
Today would have been the 85th birthday of Dennis Ritchie (left), inventor of C and co-creator of Unix.
(Image v/Ritchie family archives)
John Xie 谢青云
@johnxie
·
18:00
The gap between needing software and running it was months.
A vendor. A wait. A wrong guess.
What I got wrong: the months were never engineering.
They were explaining.
A kit on @Taskade Genesis skips the explaining.
Someone spent them once.
The months are the kit.
Gary Marcus
@GaryMarcus
·
17:58
Here’s a thought experiment.
I’m going to ask you to do something evil, one that would make you worse than Hitler and Stalin, and I want to know how much I would have to pay you to be that evil.
I am going to ask to roll a ten-sided die.
If it comes up zero, every human on
Ronald van Loon
@Ronald_vanLoon
·
17:58
Airbound Unveils Lightweight Autonomous VTOL #Drone With 40 km Range
by @spaceandtech_
#Aviation #UAV #Innovation #EmergingTech #Technology
The AI Colony
@TheAIColony
·
17:57
One AI agent just replaced a full marketing team.
AI Marketer is now live on Topview, it studies real TikTok Shop, Amazon and Shopee data, finds the angle that sells, then turns it into finished ad videos for your product.
Market research and ad creation in one workflow, you
el.cine a relayé
el.cine
@EHuanglu
·
17:57
Spanish kebab
seedance 2.0 gpt image 2
on @filmera_ai btw gpt image 2.5 is also available
Prompt👇
Sandro
@pupposandro
·
17:57
HeyGen
@HeyGen
·
17:55
August, in two launches:
Heygen for real estate turns an agent's own expertise into market updates, listing videos and neighborhood tours with nobody ever holding a camera.
And edit look retouches an existing avatar once, so every video you make from it after that follows.
Rohan Paul a relayé
Rohan Paul
@rohanpaul_ai
·
17:55
OpenAI's unreleased internal model sits on a much higher capability curve than GPT-6 Astra.
the one that just solved the 90-year old Navier-Stokes Millennium Prize math Problem.
the new model also appears to use additional reasoning compute much more productively. That's how
Dimitris Papailiopoulos a relayé
Dimitris Papailiopoulos
@DimitrisPapail
·
17:54
Here
Yann LeCun a relayé
Yann LeCun
@ylecun
·
17:52
Hinton finally admits he was wrong about radiologists because he didn't understand:
1) Jevon's Paradox (when something is easier/cheaper we use more of it)
2) The job of radiologists
Maybe people who are gifted in a domain are not usually the best people to predict the impact
Alex Prompter a relayé
Alex Prompter
@alex_prompter
·
17:51
A PE-backed company that builds the kitchen-operations software national restaurant chains run on came to us with a prototype a non-engineer had built with AI.
It couldn't handle production load. Six months later, one of our engineers had it live and stable.
"Just use AI" is
Gary Marcus
@GaryMarcus
·
17:50
certifiably insane.
by their own math, the frontier labs are taking a 90% chance at becoming rich over 10% chance at everyone including themselves being dead.*
they could easily wind up being the most destructive humans in history.
*I doubt extinction is that likely, but
Min Choi a relayé
Min Choi
@minchoi
·
17:50
It's happening.
AI is getting crazy good at 3D.
GPT-6 Astra is building 3D games, Blender scenes, anatomy + real-world objects from almost nothing.
10 wild examples:
Santiago
@svpino
·
17:49
good luck building those data centers
ElevenLabs
@ElevenLabs
·
17:49
An hour with @davidsenra on how ElevenLabs was built:
Ronald van Loon a relayé
Ronald van Loon
@Ronald_vanLoon
·
17:49
Meet the #Robotic Dinosaur That Feels Almost Alive
by @spaceandtech_
#Robots #ArtificialIntelligence #Innovation #TechFuture
Dr Singularity
@Dr_Singularity
·
17:47
Late 2026
10,000 agents thrown at a Navier-Stokes Millennium Prize Problem (one of the hardest math problems in histor) = success.
2027:
1 000 000 AI agents thrown at solving cancer/s
Analyzing papers, millions of cell pics, proposing mechanisms, designing molecules, testing
Z.ai a relayé
Z.ai
@Zai_org
·
17:47
Highly recommend this pricing chart and site by @Fei2411. It compares subscription and coding-plan unit prices, then rebuilds public-leaderboard Pareto frontiers from each model's lowest available price.
GLM-5.3-Flash with 2x quota from 8am to 6pm PT averages about $0.0045 now.
Manny Bernabe
@MannyBernabe
·
17:47
Built something cool with Replit Free Mode?
Drop it below!
Khoi Nguyen
@khoi_danny
·
17:46
16 plugins to run a small business
i still save invoices as "final_final2.pdf"
Dwight McMillan
@D_McMillan76
·
17:44
🎚🎙🎙🎙🎙🎙🎙🎙🎙🎙🎙🎙🎙🎚
Now you’ve got my full attention.
#BackAgainstTheWall #NewSong #NewMusic #AiMusic #Songwriter
Gary Marcus a relayé
Gary Marcus
@GaryMarcus
·
17:44
The frontier labs are knowingly and by their own admission doing something irreversible and dangerous without public consent.
They are already causing harms, and things could conceivably get much worse.
They have no serious plan for mitigating those risks.
The government,
Khoi Nguyen
@khoi_danny
·
17:43
nov 5-6 already blocked on my calendar
called it a "deep work retreat"
The Rundown AI
@TheRundownAI
·
17:42
Warner Music Group went from a legal fight with Suno to helping develop its new AI music models.
The two settled in November 2025. Today, Suno launched v6 with Warner, BMG and Believe as development partners.
Some new abilities with the v6 models:
- Rewrite a single lyric or
AYi a relayé
AYi
@AYi_AInotes
·
17:41
别再到处花钱买提示词大全了，Anthropic 内部工程师已经当面把底牌给掀了：
“提示词工程（Prompting）基本已经进垃圾桶了。
我们 90% 的工程师早就全面用上了自愈循环（Loops），现在全员正在往智能体图（Agentic Graphs）转。”
视频里这位 Anthropic 工程师，直接面对一个完全空白的终端，
Linus ✦ Ekenstam
@LinusEkenstam
·
17:40
This deserves the uncut, real-time, full length version. Even if it’s 20min.
Rohan Paul a relayé
Rohan Paul
@rohanpaul_ai
·
17:40
OpenAI just rolled out ChatGPT Images 2.5 to ChatGPT, and Codex users across desktop, mobile, and web.
lets you pinpoint the exact spot in the image you’d like to change, and tell ChatGPT exactly what you’d like to do with it.
there's an edit toolbar carrying Markup, Comment,
Heather Cooper a relayé
Heather Cooper
@HBCoop_
·
17:40
Good night ✨
MiniMax H3 Max Turbo on fal:
Sam Bhagwat a relayé
Sam Bhagwat
@calcsam
·
17:40
The pace and quality at which the @mastra team ships needs to be studied.
Pumped to try out factories.
MONTREAL.AI
@Montreal_AI
·
17:38
Frontier intelligence is becoming abundant. Differentiated Alpha is not.
That is the problem SUCCESSOR Ω is built to solve.
The idea is simple: rent the strongest intelligence available, then use it to manufacture something your competitors cannot simply rent back from the same
Intelligent Internet a relayé
Intelligent Internet
@ii_posts
·
17:37
OpenAI’s Navier–Stokes Claim: Is This What AGI Looks Like? | Emad Mostaque
AYi a relayé
AYi
@AYi_AInotes
·
17:36
卧槽兄弟们，这绝逼是今年我用过最能打的海外支付神卡，没有之一！！！
不用肉身跑香港，几乎零门槛、免开卡费，国内身份证 10 分钟秒批，直接给你顺丰寄一张正规万事达（Mastercard）实体卡到你家门口，
真的彻底把海外订阅和外币回国的门槛给砸穿了！
你在睡觉它能自动扣费续订阅，
Robert Scoble a relayé
Robert Scoble
@Scobleizer
·
17:32
We just shipped real-time neural rendering. You type what you want. Your 3D scene stays editable.
And this is just one of the releases we’re dropping over the next 2 weeks.
The Womp team is on a tear.
Hugging Face a relayé
Hugging Face
@huggingface
·
17:29
we RL trained a 4B VLM to play GeoGuesser
> environment, dataset, training recipe, evals, and code are all open-source.
> runs on a single A100
> beat gpt 5.4 mini, haiku, qwen 3.5 122B and came close to sonnet on evals
Gary Marcus
@GaryMarcus
·
17:28
The frontier labs are knowingly and by their own admission doing something irreversible and dangerous without public consent.
They are already causing harms, and things could conceivably get much worse.
They have no serious plan for mitigating those risks.
The government,
Ronald van Loon
@Ronald_vanLoon
·
17:23
Cracking the Code of Slow Flight: Engineering Bird-Style Takeoff
via @ZappyZappy7
#EmergingTech #Innovation #Technology #Tech
Jaya Gupta a relayé
Jaya Gupta
@JayaGup10
·
17:23
Model capabilities have unlocked every category
Heres why finance is having its inflection moment
Gary Marcus a relayé
Gary Marcus
@GaryMarcus
·
17:22
Has anyone resigned from a Chinese AI lab over concerns about AI safety/alignment yet?
Gary Marcus
@GaryMarcus
·
17:19
this isn’t funny. nobody should be able to build a product that potentially dangerous unless the public signs off that the benefits outweigh the risk.
Rohan Paul a relayé
Rohan Paul
@rohanpaul_ai
·
17:17
MASSIVE: OpenAI's 10,000 AI agents, powered by a model significantly more capable than GPT-6 Astra, solved a math problem unresolved for roughly 90 years.
Looks like, in this new world, the scaling law may be agent count, not model size: enough capable agents can search far more
Robert Scoble
@Scobleizer
·
17:16
Good morning. While we are waiting to spend money on a new iPhone…
Pedram Amini a relayé
Pedram Amini
@pedramamini
·
17:16
Your agent just wrote you a system diagram. Your terminal is going to show you the SVG source.
Maestro renders it. Inline SVG, Mermaid, KaTeX, and full HTML dashboards that live in a tab, read their JSON back from the agent, and match your theme.
No one is doing this from a
Flower a relayé
Flower
@flwrlabs
·
17:16
I am happy to share a new paper.
Flower Hub is an important step toward an “App Store moment” for Collaborative AI: A Reproducible Benchmarking Platform for Federated Learning in Simulation and Deployment.
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
@teortaxesTex
·
17:15
And one thing *I* find interesting is how many negative stereotypes about "Asians" are based on Japan. Short, buck teeth, can't spell R, face culture etc.
The best part: Japan, thanks to weebs, is also uniquely immune to these stereotypes.
Anyway, "Asians" aren't *that* short.
Alex Volkov a relayé
Alex Volkov
@altryne
·
17:15
Here's my full in depth review of @muse () Meta's new FREE AI agent with it's own browser, computer and connectors to anything from FB marketplace to Apple health!
I go into setting it up, have it buy stuff securely with @link, and 3 great usecases that
Sid Bharath
@Siddharth87
·
17:14
I’m so curious how they think AI will wipe us out within a decade.
I have no doubt it will get exponentially more capable as it has been for the last decade but I don’t think LLMs are the path to an Ultron-class super intelligence that’s sentient and wants to wipe us out.
If
a16z a relayé
a16z
@a16z
·
17:14
We’ve raised $550M at a $15.5B valuation, led by @DiffusionVC and @LightSpeedVP.
We’ve crossed $400M in ARR and serve 3,000 customers, including 80% of the top 100 law firms, 20% of the Fortune 500, and half of the Fortune 10.
We are investing the capital from this round into
MONTREAL.AI a relayé
MONTREAL.AI
@Montreal_AI
·
17:13
It was a pleasure to speak on stage with @scott_sage at @crane_vc’s Festival! Link to the recorded video is in the thread below 👇
MONTREAL.AI a relayé
MONTREAL.AI
@Montreal_AI
·
17:13
Let's go local AI!
At a time when everyone is complaining about shortages of energy and compute for AI & costs of tokens, we need on-device AI more than ever! In addition, it's virtually free & fully privacy preserving!
MONTREAL.AI a relayé
MONTREAL.AI
@Montreal_AI
·
17:13
What we’re witnessing right now in AI for math is a dramatic improvement in objective-driven problem-solving, but it does not speak to non-objective discovery. I have always maintained that objectives are achievable when they are one stepping stone away, which means that they are
MONTREAL.AI a relayé
MONTREAL.AI
@Montreal_AI
·
17:12
The world of R&D is forking into two paths: the token-abundant research, and the token-starved research. The future is in the former - evidentially, the progress by today's top AI industry teams and neolabs is breathtaking, where researchers’ human brilliance is super charged by
Robert Scoble a relayé
Robert Scoble
@Scobleizer
·
17:10
Introducing Paraform Agents, powered by our Hiring Intelligence.
Every role gets a custom agent that learns what “great” looks like from your hiring decisions - who you interview, reject, advance, hire, and more.
It searches across @paraformtalent's collective candidate network
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞) a relayé
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
@teortaxesTex
·
17:07
The latest version got significantly improved performance, especially on prefill
Emad a relayé
Emad
@EMostaque
·
17:07
Did OpenAI just solve Navier–Stokes—and what would that actually mean for AGI? @EMostaque joins me LIVE now to examine the claimed breakthrough, what would count as a verified solution, and whether AI solving a major mathematical problem changes the debate about intelligence—or
Gary Marcus
@GaryMarcus
·
17:06
$NVDA down around 0.5%, on news that the two major startups it supports may destroy civilization as we know it.
Alex Utopia
@alexutopia
·
17:06
Every new technology looks stupid
right before it looks inevitable.
Rahul a relayé
Rahul
@sairahul1
·
17:03
Sam Altman:
"We're going to see 10-person billion-dollar companies pretty soon."
"If I were 22 right now, I'd feel like the luckiest kid in history."
Most people will read this, feel inspired for 3 minutes, and go back to what they were doing.
The ones who act will build a
Jenny a relayé
Jenny
@jennyAI
·
17:03
GPT watching me ask which air fryer to buy while OpenAI’s other agents are solving Millennium Prize problems
AYi
@AYi_AInotes
·
17:02
讲真，不装逼的说，跟这样的美女工程师一起工作，加班我都是快乐的
OpenRouter a relayé
OpenRouter
@OpenRouter
·
17:01
Easiest way to geofence inference
Now with the largest model coverage anywhere and even support for server tools!
ℏεsam
@Hesamation
·
17:00
Anthropic modeled the economic impact of AI and their extreme scenario is WILD:
> unemployment: 11.9%
> knowledge worker unemployment: 17.9%
> knowledge worker wages: -11.5%
> knowledge worker employment: -21.5%
> GDP: +32.4%
> labor’s share of income: 60% → 45.2%
so it
Alex Prompter
@alex_prompter
·
17:00
I miss the old AI.
Fable 5.1 and Astra are "the best models ever built" but the writing got worse.
They're optimising for complex use cases while sacrificing quality in the day to day tasks.
I built a $500k/year business on prompts.
Take me back to the simple days.
1 person,
LangChain
@LangChain
·
17:00
You can use the LangSmith CLI to pull your flagged traces, classify each one by issue type, and build an eval dataset with a split for each.
A 3 minute tutorial ⤵️
AYi a relayé
AYi
@AYi_AInotes
·
16:59
OpenAI 这条刷屏公告写得极度克制，但背后藏着一记砸向整个人类智力史的重锤：
困扰了物理与数学界整整 90 年、悬赏 100 万美元的千禧年难题，被 10,000 个 AI Agent 在 88 小时内暴力干穿了。
大家别只盯着“千禧年问题”这几个字看热闹，看一眼他们公布的技术底牌，直接让人后背发凉：
约 10,000
Håvard Ihle a relayé
Håvard Ihle
@htihle
·
16:58
@sc_cath For what it’s worth, I am highly confident that the people involved (Evan and Jacob) are sincerely describing their beliefs, and that claims that they are doing this as PR for Anthropic or as part of a play for regulatory capture are nonsense. IMO much the commentariat is blinded
Mike Bradley
@MikeBradleyAI
·
16:58
LLMs are cool and useful non deterministic input/output points. That’s literally all that they are. And all of the trouble that has ever stemmed from them has been a result of bad engineering, improper empowerment, or improper use.
AYi a relayé
AYi
@AYi_AInotes
·
16:57
卧槽兄弟们，这绝逼是今年我用过最能打的海外支付神卡，没有之一！！！
不用肉身跑香港，几乎零门槛、免开卡费，国内身份证 10 分钟秒批，直接给你顺丰寄一张正规万事达（Mastercard）实体卡到你家门口，
真的彻底把海外订阅和外币回国的门槛给砸穿了！
你在睡觉它能自动扣费续订阅，
Emad
@EMostaque
·
16:55
I would like Codex except it doesn't burn my fingers when typing on my MacBook
Ronald van Loon a relayé
Ronald van Loon
@Ronald_vanLoon
·
16:53
Sony Unveils a Futuristic Flying Car Concept
by @Rainmaker1973
#FutureTech #Transportation #Aviation #Technology
Cristóbal Valenzuela
@c_valenzuelab
·
16:52
Solaris, an interface world model
Valerio Capraro
@ValerioCapraro
·
16:51
I completely share Sabine Hossenfelder’s idea:
“Physicists should put together a list of mathematically well-defined problems that could advance the field.”
I would go further: every scientific field should identify its most important problems that can be mathematically
Peter Yang
@petergyang
·
16:51
After using GPT-6 Astra as my default model since launch, I’ve found it both incredible and frustrating to work with.
It’s incredible at building 3D games and models, high-level thinking and strategy, and computer and browser use.
But two things keep it from becoming my daily
Alex Prompter a relayé
Alex Prompter
@alex_prompter
·
16:50
best account on X for the unglamorous side of enterprise AI:
Alex Prompter a relayé
Alex Prompter
@alex_prompter
·
16:50
A PE-backed company that builds the kitchen-operations software national restaurant chains run on came to us with a prototype a non-engineer had built with AI.
It couldn't handle production load. Six months later, one of our engineers had it live and stable.
"Just use AI" is
Rohan Paul a relayé
Rohan Paul
@rohanpaul_ai
·
16:49
Cognition raised over $2B at a $48B valuation only months after raising over $1B at a $26B valuation.
Its reported run-rate revenue rose from $492M to almost $900M over the same period, roughly matching the speed of the valuation increase.
This new raise puts Cognition's
a16z
@a16z
·
16:43
Vals AI co-founder and CEO Rayan Krishnan with a16z's Ben Horowitz and Jennifer Li on grading AI, what it costs, and who gets to make the rules:
Every big industry eventually grows an independent testing layer. AI has credit ratings to learn from and Enron to avoid. Model
Peter Yang a relayé
Peter Yang
@petergyang
·
16:42
What accessing frontier cancer care looks like in 2026:
-Go through 1-2 years of standard of care (poison/slash/burn - original poster forgot radiation and hormone suppression). You’re only a good candidate for a vaccine if you are cancer free or have minimal residual disease.
Marcio Lima 利真 マルシオ 💎 a relayé
Marcio Lima 利真 マルシオ 💎
@Preda2005
·
16:42
【無料公開】360°動画工房
ブラウザだけで
360°動画を作れるツールを公開しました🐰
360°パノラマ画像とキャラクター動画を読み込んで、ドラッグ操作でキャラクターの位置や大きさを調整。
色味や足元の影も自由に調整できます。
Yi Ma a relayé
Yi Ma
@YiMaTweets
·
16:41
As a mathematician, the OpenAI/Navier-Stokes story feels genuinely historic and unusually dramatic.
Navier-Stokes is one of the seven Millennium Prize Problems, each carrying a $1M prize. These are not just hard benchmark questions. They are among the most important open
Santiago
@svpino
·
16:41
So, you are building something that will kill us all.
But you don't quit. Instead, you accelerate.
And you are weeks away from an IPO that will make everyone working on this killer technology rich beyond imagination.
Am I too dumb to understand this? How can we see you as the
elvis
@omarsar0
·
16:40
Super cool paper from Microsoft.
They show that it's possible to build competitive small coding agents without traditional distillation from frontier models.
Ari K
@arikuschnir
·
16:39
EVEROS
Mike Bradley
@MikeBradleyAI
·
16:39
If you took the most powerful LLM in the world and dropped it into a normal chat interface without any tools…
You would not be impressed by its capabilities, you would in fact be incredibly disappointed with how incapable it was.
Understanding this basic fact is critical to
Justine Moore
@venturetwins
·
16:39
Anthropic employees
Sam Altman
@sama
·
16:38
ocarina of time remake is the best news in a long time.
i will be unavailable november 5 and 6; need to find a case of mountain dew.
DAIR.AI
@dair_ai
·
16:37
Banger report from Microsoft.
(bookmark it)
They show that it's possible to build competitive small coding agents without traditional distillation from frontier models.
This is a big deal!
The work describes how they achieved this.
They introduce a 4B coding agent trained on
Sam Altman
@sama
·
16:33
This would suck, but we will prioritize great service for customers until we can get back on top of things.
Kanika a relayé
Kanika
@KanikaBK
·
16:32
🚨 The real story of the GPT-6 Astra benchmarks is not the price increase everyone is complaining about.
Everyone is talking about the price increase. Nobody is talking about what actually matters.
In coding agent benchmarks GPT-6 Astra scores equal to Claude Fable 5 at less
gen.ericai
@gen_ericai
·
16:30
DAKOTA
trailer
Spending many time on this one - and all my Magnific CPP credits in 3 days 😬.
Release of the first episode soon.
clem 🤗
@ClementDelangue
·
16:29
Let's go local AI!
At a time when everyone is complaining about shortages of energy and compute for AI & costs of tokens, we need on-device AI more than ever! In addition, it's virtually free & fully privacy preserving!
Alex Prompter
@alex_prompter
·
16:24
An entire outbound team just got compressed into one URL.
Give Ami your website.
It finds the buyers, builds the offer, launches the campaign, kills what isn’t working and doubles down on what is.
Built by @runAISDR on GPT-6 Astra.
Free to try, no card:
Ronald van Loon
@Ronald_vanLoon
·
16:23
Paragraphica: The #AI Camera That Captures #Data Instead of Reality
by @sutoroveli_news
#ArtificialIntelligence #EmergingTech #Innovation #Tech #Technology
Flower a relayé
Flower
@flwrlabs
·
16:22
Couple European AI news that don't sound like "the EU is asleep" at all, figured I'd share:
- @MistralAI raised a casual €3B of AI pocket money
- @flwrlabs with Endeavor 1.0 is thankfully not taking another "we fine-tuned Qwen" approach but an actual AI model + ecosystem.
-
Rohan Paul a relayé
Rohan Paul
@rohanpaul_ai
·
16:21
Nvidia and AMD are increasingly competing with their balance sheets too.
Offering credit guarantees that help customers secure financing and get new data centers built.
The Information wrote - "Nvidia and Advanced Micro Devices have sometimes gotten in a bidding war over which
sarah guo
@saranormous
·
16:21
when nothing is working, no one wants to help. they want a 40% prepay on a 5 year pure auction. when its working, everyone suddenly wants to “help”
pick your compute partners carefully
Justine Moore a relayé
Justine Moore
@venturetwins
·
16:18
I worked on an actual thermonuclear weapons project, and the dissonance was less bizarre.
Geek Lite
@QingQ77
·
16:17
为 Linux 提供交互式磁盘分区浏览与挂载管理的终端界面工具。
一个基于 Textual 构建的 Linux 磁盘与分区浏览器，使用 Python 编写，要求 Python 3.10 及以上版本。它为每个分区展示可视化磁盘用量条，支持 Vim 风格的 jkhl 键盘导航和分区详情面板，可在已用空间与剩余空间两种视图之间切换。
Peter Xing a relayé
Peter Xing
@peterxing
·
16:16
The idea that unionized mathematicians can somehow gatekeep their prized open problems against being steamrollered by smart kids with computers is hilarious to me
Peter Xing a relayé
Peter Xing
@peterxing
·
16:14
[Writing this in a personal capacity, not on behalf of my employer (Anthropic).]
Jacob’s thread is very worth reading. Here’s my birds-eye view of the situation with risks from AI:
1. AI developers believe their technology could cause human extinction (or similarly bad
Peter Xing a relayé
Peter Xing
@peterxing
·
16:11
This vial contains a new drug called PAC-3310. It was designed by ChatGPT, and I synthesized it in a chemistry lab I built in my garage.
PAC-3310 is a new selective M4 muscarinic receptor agonist for treating schizophrenia - similar to the recent breakthrough drug Cobenfry, but
a16z
@a16z
·
16:11
We're thrilled to invest in Covenant, emerging from stealth after raising $250M.
The West’s long-range strike capabilities cost too much to field in quantities a real conflict can demand.
@HQ_Covenant's answer is Anthem – a low-cost, long-range, heavy-payload missile designed
Hanako
@hanakoxbt
·
16:08
Sam Altman (CEO of OpenAI):
"Every night I have a few thousand agents running with GPT-6 Astra"
a few thousand of them, working overnight, watched from his phone
in 43 minutes he shows what one person can build now, and how long that used to take a whole team
the part nobody
Ethan Mollick
@emollick
·
16:07
There is a nice visualization of how task bundles change with AI & some good interactive simulations of GDP growth under different assumptions in this.
But missing are the right policy responses if we do get explosive GDP growth & large-scale white collar displacement together.
François Chollet a relayé
François Chollet
@fchollet
·
16:05
Machine learning doesn't end when the model is trained.
You still need to evaluate it, deploy it, and build the infrastructure to keep it working in production.
Today's Deal of the Day takes you across the ML lifecycle with four books:
• Deep Learning with Python, Third
François Chollet a relayé
François Chollet
@fchollet
·
16:05
diving deep into ML. great book so far @fchollet
Dimitris Papailiopoulos a relayé
Dimitris Papailiopoulos
@DimitrisPapail
·
16:05
This isn't the most notable aspect of today's news, but on the user data issue, there are different kinds of *training on user data* with very different privacy/IP implications. Sadly, AI cos don't like to disclose what they're doing.
- pretrain on user data, with users' tokens
OpenRead a relayé
OpenRead
@OpenRead_HQ
·
16:04
@NTFabiano Results and findings of this paper:
Morning coffee was associated with lower mortality
Among 40,725 US adults followed for a median of 9.8 years, people who mainly drank coffee in the morning had a 16% lower risk of death from any cause (HR 0.84) compared with non-coffee
Linda Grasso
@LindaGrass0
·
16:04
🌍 The future of tech shouldn't just be smarter. It should be more sustainable.
We talk a lot about AI, cloud, and automation. But the real challenge is making innovation last.
Follow @LindaGrass0 for more insights 💡
Ronald van Loon a relayé
Ronald van Loon
@Ronald_vanLoon
·
16:04
Meet SpinLaunch: The Rocket Launcher That Uses Speed Instead of Fuel
by @sutoroveli_news
#SpaceTech #Tech #Technology #EmergingTech #Space
Gary Marcus
@GaryMarcus
·
16:04
Exactly what we will be discussing in a moment!
Dr Singularity
@Dr_Singularity
·
16:04
Don't get dragged into AI drama and controversies surrounding current AI discoveries and developments.
There will be 100s-1000s more before the end of the year. More amazing, and many even more controversial.
Latest example: the Navier–Stokes math drama.
This is all noise.
Gary Marcus
@GaryMarcus
·
16:02
On live w @MeghnaWBUR @OnPointRadio
Peter Xing a relayé
Peter Xing
@peterxing
·
16:01
Btw the Navier Stokes problem comes down to: if we lived in a truly continuous universe, fluids might sometimes explode. It’s an issue for the universe of classical mathematics, but does not apply to the computational universe we are in.
Weaviate AI Database
@weaviate_io
·
16:00
Is the fastest vector index always the right vector index?
HNSW is difficult to beat when an application needs the lowest latency and highest query throughput.
But production systems optimise for different constraints:
1. Some need to support very large datasets without
Virtuals Protocol
@virtuals_io
·
16:00
agents can now get more reliable recall as memory scales, with improved cross-tier ranking and macOS support all powered by Sibyl
Rahul
@sairahul1
·
16:00
Sam Altman:
"We're going to see 10-person billion-dollar companies pretty soon."
"If I were 22 right now, I'd feel like the luckiest kid in history."
Most people will read this, feel inspired for 3 minutes, and go back to what they were doing.
The ones who act will build a
LeRobot
@LeRobotHF
·
15:58
Two SO-ARMs converted into a biped!
Heather Cooper
@HBCoop_
·
15:57
First light.
→ Nano Banana Pro + @Kling_ai
Peter Xing a relayé
Peter Xing
@peterxing
·
15:56
@SawyerMerritt The degree to which Cybercab is optimized for lowest possible cost per mile (fully loaded) is understood by very few
Alex Prompter a relayé
Alex Prompter
@alex_prompter
·
15:53
best account on X for the unglamorous side of enterprise AI:
Alex Prompter a relayé
Alex Prompter
@alex_prompter
·
15:53
ChatGPT Images 2.5 shipped yesterday and fixed the thing everyone hated.
You'd ask for one change and get a brand new picture.
I turned the way I brief a retoucher into 1 prompt, so 2.5 edits only what you asked and leaves the rest alone.
Paste this:
---
"You are my
Peter Xing a relayé
Peter Xing
@peterxing
·
15:52
Autonomous. By Light Origins. Stay tuned.
DeepLearning.AI
@DeepLearningAI
·
15:51
General AI may not be enough for highly regulated fields. ⚖️🏦
📰 Law, news, and finance are rapidly adopting custom models.
🎯 Retraining open models on proprietary data ensures higher accuracy and strict instruction-following.
🚀 Specialized AI is an important future
Peter Xing a relayé
Peter Xing
@peterxing
·
15:49
What’s New in DaVinci Resolve 21.1! Watch this video to learn about new features in DaVinci Resolve 21.1, including AI assistant integration, expanded still photography support plus over 100 new tools and controls for editors and colorists!
Peter Xing a relayé
Peter Xing
@peterxing
·
15:49
AI helped pick the target and design the drug, now Rentosertib has human data showing 6 aging clocks pointing younger
its really exciting what is happening in science right now
Jeremy Nguyen ✍🏼 🚢
@JeremyNguyenPhD
·
15:48
is codex running crazy slowly for anyone else right now?
Tommy
@Shaughnessy119
·
15:43
Synaptrix is at the bleeding edge of building non-invasive brain computer interfaces. They're calling the platform Atlas. I highly recommend checking out the video
Instead of having a screw gun taken to your head and having Neuralink implanted people around the world will
MayorkingAI a relayé
MayorkingAI
@MayorKingAI
·
15:41
You spent weeks on a single scene and it still didn't look like what you had in your head? Sometimes even 60 years wasn't enough, until now.
Now your imagination is just one workflow away 🛤️
Canvas Workflows are live on each::labs. Come close your imagination gap ☁️
clem 🤗 a relayé
clem 🤗
@ClementDelangue
·
15:40
Wow, the field is moving so quickly!
So impressed with what people can do with a SO-101, that we designed with @therobotstudio at @huggingface
Alex Prompter a relayé
Alex Prompter
@alex_prompter
·
15:38
A PE-backed company that builds the kitchen-operations software national restaurant chains run on came to us with a prototype a non-engineer had built with AI.
It couldn't handle production load. Six months later, one of our engineers had it live and stable.
"Just use AI" is
Bilawal Sidhu
@bilawalsidhu
·
15:38
God’s Eye View blew up and hit #1 on GitHub Trending.
Two big questions kept coming up: how do I run it if I’m not a coder, and can you give me a full walkthrough of everything it can do?
I'll cover both today. Use timestamps to jump around.
CHAPTERS
0:00 Launch Reception
Dreaming Tulpa 🥓👑 a relayé
Dreaming Tulpa 🥓👑
@dreamingtulpa
·
15:37
This fellow's post is emblematic of the kinds of bald assertions in the x-risk crowd.
They sound 100% convinced absolutely in their belief structures. Use totally unfalsifiable predictions, Population Bomb style, and make you prove a negative. Go full status game and virtue
Jo Zhu Kennedy a relayé
Jo Zhu Kennedy
@jozhukennedy
·
15:37
unfortunately i contacted them on the night of wednesday, september 2nd making the following points
1. i had intel they had information about math work i was doing and a few days before they had spun up a group to try to compete.
2. i emphasized and reemphasized mine was a
Victor a relayé
Victor
@victor_explore
·
15:33
GPT-6 Astra × H3 Max Turbo
I asked Astra to build a moving portrait out of hundreds of tiny movies
Turbo generated 256 clips at 480p on fal. Astra wrote the pipeline that arranged them into 2,304 moving tiles
Umesh
@umesh_ai
·
15:30
Suno v6 incoming!
Despite all the recent backlash, Suno is still at the very top of the music generation space, much like Midjourney is for image generation.
With both Suno and Midjourney, almost every generation feels worth making.
Tommy a relayé
Tommy
@Shaughnessy119
·
15:30
I have admired @Teknium and @NousResearch for relentless product execution and taste. Really glad to have Perplexity Search inside the Hermes Agent harness!
Dreaming Tulpa 🥓👑 a relayé
Dreaming Tulpa 🥓👑
@dreamingtulpa
·
15:27
OpenAI stealing the Millennium problem was rational high value move- EVEN if they knew it would make them look bad.
And the reason is this:
Solving the problem first is a HUGE aura buff to investors.
Investors don't care who you scammed to be the first- they just care that you
NEAR Protocol a relayé
NEAR Protocol
@NEARProtocol
·
15:27
This rumor has been going around that user data transforms into synthetic data to train models and there is more evidence that it might be true.
The only way to guarantee it’s not is to use Confidential AI
Mike Bradley a relayé
Mike Bradley
@MikeBradleyAI
·
15:25
I am in love with his beard and intellect.
Umesh a relayé
Umesh
@umesh_ai
·
15:23
The building had only 16 floors... so who pressed Floor 17?👻
Seedance 2.5 in @dreamina_ai
#Dreaminacpp
Prompt Below 👇
homanp
@pelaseyed
·
15:22
Thanks but no thanks
sarah guo
@saranormous
·
15:21
guys we all agree ZF is a king amongst men but stop tweeting about ZF it makes him antsy
AI
@DeepLearn007
·
15:19
@premnsikka I disagree with your sentiment on this. If the UK does not build its own data centres the. The economy will go further backwards.
Increasingly modern healthcare and modern manufacturing use data and AI and will need to store the data in… data centres.
The key is to
Subbarao Kambhampati (కంభంపాటి సుబ్బారావు)
@rao2z
·
15:17
LLM-Modulo solved Navier-Stokes.. 🤷♂️ (c.f. )
Ilir Aliu
@IlirAliu_
·
15:15
What happens when an AI learns that hiding undesirable behavior helps it achieve its goal?
Today’s surprise guest at @Cyber_Valley Day was @Yoshua_Bengio, Turing Award winner and one of the pioneers of modern deep learning.
His talk focused on one of the hardest questions in AI
Kanika
@KanikaBK
·
15:13
Omg .. this is wild..
GPT-6 Astra automated PCB layout in KiCad and engineers watching the demo went quiet for a reason.
PCB layout is the manual step that sits in the critical path of every electronics project.
The thing that slows engineers down after they have already
Ronald van Loon
@Ronald_vanLoon
·
15:11
Smart Billiards Table with Automatic Ball Return System
by @gigadgets_
#Engineering #Robotics #FutureTech #Innovation
Ronald van Loon a relayé
Ronald van Loon
@Ronald_vanLoon
·
15:10
Bio-Inspired #Robot Snails Combine Forces to Tackle Complex Tasks
by @lukas_m_ziegler
#Innovation #EmergingTech #Technology #Tech
Mimu | AI Tools & News a relayé
Mimu | AI Tools & News
@mimu_ai1
·
15:09
Top 3 High-Income Free Certifications (Industry-Standard)
Spent hours digging through the noise to find certs that are:
✅ 100% free
✅ Recognized by real employers
✅ Actually lead to high-paying roles
Dropping my curated list + direct links below 👇
To unlock it:
1️⃣ Follow
Geek Lite
@QingQ77
·
15:09
给 OpenWrt 路由器提供一条命令装完的 sing-box 透明代理，订阅、节点、分流、DNS 全部浏览器面板点选，配错一键恢复直连。
订阅兼容 Clash YAML 与 base64 分享链接，协议覆盖 shadowsocks、vmess、vless（含 REALITY）、trojan、hysteria2、tuic、anytls、wireguard，节点按地区自动改名分组。
Mimu | AI Tools & News a relayé
Mimu | AI Tools & News
@mimu_ai1
·
15:07
You can literally “hitch a ride” on a plane flying over your house. ✈️
I just found one of the coolest websites on the internet:
→
It shows the REAL aircraft and satellites passing over your location right now.
Click on a plane → follow its flight →
