News

Companies Funds News People Jobs Events Mafias Charts Map Money

Loading the latest news

The wire.

The latest stories and conversations from New York tech.

items

sources

mentioned

Follow a topic

In the conversation

Mentions

Where it comes from

Items

NYTech · the New York tech scene, citedMCP Hendrik De Winne Corrections

The wire.

The latest stories and conversations from New York tech.

Refreshed hourly · Latest 3:52 PM ET

120items

6sources ↘89mentioned ↘

Sunday, October 4

10 items

Alex Wiltschkoreposted3:52 PM

RepostedCristóbal Valenzuela@c_valenzuelabp(spark) is the probability that AI becomes fully creative. Fully creative means that AI makes new ideas, styles, or forms of art on its own, without human intervention or human examples at the level of the best human creators. My current p(spark) is ~90%.

Michael Dempseyvia @mhdempsey1:54 PM

There are pivotal times in the company building journey where you must light a fire. Sometimes that means burning a bridge so there's no going back. Other times it means burning down what you've built and starting over. We have found @compoundvc that sometimes founders need https://t.co/8jGCj0NtLV

Attached image 1

Attached image 2

Anthropicvia AlleyWatch1:43 PMLatest story

Should AI Be Regulated City by City? New York Should Think Carefully Before Answering

New York has real reasons to worry about AI. But a city-run permit system for every AI tool sold here would hit startups hardest.

Alex Wiltschkoreposted7:01 AM

RepostedShahab Bakhtiari@ShahabBakhtI had a feeling recently that seems to be corroborated by chats with colleagues, so I decided to share it here. A lot of systems neuroscientists driven by algorithmic-level questions are going through what looks and feels like a research depression. 🧵 (1/n)

Thomas Wolfvia @Thom_Wolf4:08 AM

new roommate just moved in. walks like he's had three drinks, says no to everything. still quite cute so I might bring him everywhere with me https://t.co/xT4oBBrrVs

Attached image 1

Max snowreposted12:10 AM

RepostedLaura Loomer@LauraLoomerAI officials and several Trump administration officials caught with their pants down hanging out at Harvard 2 days ago with Jake AutoPen Sullivan and Susan Rice. Does everyone remember what Sullivan and Rice did to President Trump? Interesting. Why were the people on this list at a closed door secret summit the same time the AI execs were at the White House kissing President Trump’s ass pretending like they don’t hate him and the GOP? 🤔 Something smells like shit. Sullivan and Rice belong in GITMO. They should have been indicted years ago! @AGToddBlanche

Max snowreposted12:09 AM

RepostedEnd Wokeness@EndWokeness91% of black female homicides in the US are committed by black men 8% are committed by black women 1% from all other categories

Max snowreposted12:07 AM

RepostedLibs of TikTok@libsoftiktokOMG. Adam Quach, a MAN who goes by the name "Alice" and uses "She/her" pronouns, was hired to coach the Etiwanda High School (@CjuhsdS) girls' volleyball team in CA. He was just ARRESTED on charges of Possession of child p*rnography. Are you surprised? https://t.co/4MVmZrCJ1C

Max snowreposted12:07 AM

RepostedLaura Loomer@LauraLoomerThe Islamic takeover of America continues. My friend is in San Diego right now at Navy Pier in front of the USS Midway Carrier for a birthday weekend getaway and they just sent me these photos of Muslims trying to convert American military members to Islam at the Navy Pier. The Muslims are walking up to US Navy members trying to convert them to Islam. They even have a sign that says “Jesus Is a Muslim”. Their tent in front of the military carrier is giving away free Qurans in English and Spanish!!!! It’s interesting these Muslims are giving out free Qurans in English and Spanish in San Diego of all places. San Diego is where a large portion of illegal aliens live because it’s on the border of Mexico. California is also a sanctuary state for illegals. It’s worth noting that this week, California Governor @GavinNewsom signed Assembly Bill 2017 into law, officially designating the Islamic animal sacrifice days of Eid al-Fitr and Eid al-Adha as state holidays in California! I hope everybody is paying attention to how much the Islamic takeover of America is escalating. This is truly disgusting.

Quoted image 1

Quoted image 2

Max snowreposted12:06 AM

RepostedHeidi Ganahl@heidiganahlColorado politicians love to call our elections the "gold standard." In 2024, we checked. It's fool's gold. 🪨 In 2020, Jena Griswold told Coloradans that Wi-Fi and Bluetooth were "stripped from the unit before use" on our voting systems. So in 2024, we started digging. Here's what our investigation found: 🗳️ Douglas County: we scanned more than 85,000 paper ballots. The ballots themselves checked out. The problems were in the system around them. 🔍 Wireless networking cards were physically installed in voting computers in 12 Colorado counties. Purchase orders and county clerks confirmed it. One clerk admitted it was "a known thing amongst the counties." 🔑 More than 600 BIOS passwords for Colorado's voting equipment sat on the Secretary of State's own website. For MONTHS. 📬 We audited undeliverable ballots in nine counties, matching the Post Office invoices that are supposed to go to county clerks against the counties' own counts. 30,000 undeliverable ballots were unaccounted for. 📋 Dirty voter rolls, full of people who shouldn't still be on them. 📹 Unmonitored drop boxes, with cameras that can't tell you who dropped off a ballot or how many. ✍️ Signature verification, the main safeguard on a mail ballot, has never been proven to work in Colorado. 🔒 And no independent audits. Nobody outside the Secretary of State's office checks any of it. The system grades its own homework. None of this came from the people in charge. It came from citizens who asked questions and wouldn't take "trust us" for an answer. That was two years ago. Ballots for 2026 are in your mailbox right now. Has anyone in charge explained any of this? And the official who told you the Wi-Fi was gone? She now wants to be your Attorney General. 👇 Our 2024 investigation: https://t.co/dvaSBGUge5 #copolitics #ElectionIntegrity @COIFFEtable

Saturday, October 3

40 items

Thomasreposted11:22 PM

Repostedilliquid@lefttailguyI wonder what's next. I suspect one interesting endpoint is that the excel model becomes the quantitative surface of a much deeper research graph / self-constructing ontology. I could imagine something like every quantitative revenue key driver being hyperlinked to a graph of the intermediate qualitative assumptions that underpin it. Maybe analysts use some kind of ambient capture during their desktop and field research process, and then there's an agent-mediated live evidence stream (e.g, expert calls, filings, transcripts, podcasts, articles, tweets, etc.) that supports or contradicts the analyst's assumptions in real time. And based on that you could calibrate your confidence about various qualitative assumptions at any given point (as new evidence comes in) and the model could compute how that flows through to quantitative KDs. Overtime, this surface / system would learn how new evidence shifts your priors and how accurately your updates actually end up representing economic reality. It could then start to understand how to weight different pieces of qualitative evidence! Lots to think about.

judy shapiroreposted10:03 PM

RepostedMoms for Liberty@Moms4LibertyParental rights start with parental awareness. Not everything made for children is as innocent, or as neutral, as it may seem. That doesn’t mean parents need to panic. It means we need to PAY ATTENTION. Know what your kids are watching. Know what they’re reading. Know what they’re being taught. Know who is influencing them. Because you can’t exercise your parental rights if you don’t know what your child is being exposed to. Ms. Rachel built an audience by teaching toddlers songs, words, and numbers. Now, her political advocacy and messaging are raising questions about whether parents have enough information to make informed decisions about what their youngest children are watching. You are the parent. You decide what is appropriate for your child. But first - you have to know. READ MORE: https://t.co/JsBs1uRFQH #moms4liberty #parentalrights

Topic Intelligence™️

clem 🤗reposted9:55 PM

RepostedBinh@pham_blnhI just made the most expressive real-time motion harness for the Reachy Mini. It was created by: > synthesizing a massive motion dataset using Astra (over 10000 samples) > fine-tuning Qwen 3.5 4B using such dataset > then training a flow matching transformer head to transform sparse motion plan (<1hz) into dense expressive motion (25Hz) The average inference time for a full motion is less than 200ms, enabling real-time interaction. @pollenrobotics and @huggingface only recorded 85 emotions for the reachy mini. With my work, you can now prompt and animate any motion you want. Everything is open-source; write-up and links attached below. Here is a video of Reachy explaining itself through my harness.

Clem Delangue PClément Delangue

judy shapiroreposted9:53 PM

RepostedDan Burmawi@DanBurmawyMost people have no idea about the real history of Al-Aqsa Mosque, because they’ve only been fed the Islamic Palestinian propaganda version of events. They believe that Al-Aqsa has always been Islam’s third holiest site, that it has belonged to Muslims since the dawn of time, and that Israel is the oppressor for merely existing near it. None of that is true. The Temple Mount in Jerusalem was the holiest site in Judaism for over a thousand years before Islam even appeared in history. It housed the First and Second Jewish Temples, the center of Jewish worship and pilgrimage. When the Romans destroyed the Second Temple in 70 AD, they built pagan shrines over it, but its Jewish identity never disappeared. In the 7th century Islam emerges and expands through conquest, and begins hijacking Jewish and Christian sites, prophets, and narratives. At first, Jerusalem had no major significance in Islam. Muhammad never set foot there. There was no mosque. There was no pilgrimage. There was no Islamic history tied to the city. But that changed during the brutal power struggle between Abd al-Malik ibn Marwan and Abdullah ibn al-Zubayr. By the late 7th century, Islam was deeply divided. Abdullah ibn al-Zubayr controlled Mecca and Medina, the two holiest cities in Islam. And Abd al-Malik ibn Marwan, the Umayyad Caliph, controlled the Levant. But Abd al-Malik had a problem, he didn’t want the people of the Levant traveling to Mecca for pilgrimage, because that would give power to his rival. Abd al-Malik declared the Temple Mount as Alaqsa mosque that was mentioned in the quran, and he made it an alternative place of pilgrimage. He ordered the construction of the Dome of the Rock and the Al-Aqsa Mosque to divert attention from Mecca. And just like that, Islam manufactured a holy site for political gain. The real Masjid Al-Aqsa referred to in the Quran was not in Jerusalem, it was between Mecca and Ta’if. There was no mosque in Jerusalem at the time. There was no Islamic presence there. Yet, centuries later, after Islam had conquered the city, the Islamic narrative retroactively applied this Quranic verse to Jerusalem, again, for political convenience. If Israel wanted to act like Islamic conquerors, it could have easily done to Al-Aqsa what Turkey did to the Hagia Sophia. It could have converted the mosque into the Third Temple, banned Muslim prayer on the site, erased any trace of Islamic history, as Muslims did to Christian and Jewish sites throughout history. But Israel didn’t do that. Israel allows Muslims to pray there freely. Israel protects Al-Aqsa, even as it is used to spread anti-Semitic propaganda and incite violence. Yet, despite this, the world condemns Israel for merely existing in its own capital.

Topic Intelligence™️

nihalreposted7:52 PM

Repostedsunny madra@sundeepSometimes you gotta have hard conversations with these agents https://t.co/cMCgqPV0Wa

Quoted image 1

Quoted image 2

Cory Levyvia @cory6:18 PM

visiting Harvard and MIT soon @ reply or DM if ur there!

wayne nelmsvia @wayne_nelmz4:18 PM

All of the investment grade compute deals seem to be done. Who will be the first AI cloud and DC providers that will be comfortable underwriting the future of non-investment grade offtake?

Michael Mignanovia @mignano3:30 PM

A bunch of you asked me who produced this video. It was @sandwich. They are world class. 10/10 on both creative and production, would work with them again, every time.

QuotedMichael Mignano@mignanoIntroducing @Supertake, a new platform that transforms your unique takes on the world into real, shareable investment portfolios using frontier AI and trading agents. Supertake makes it easy for anyone to invest in their ideas, even if they know nothing about investing. 1/ https://t.co/pgettfU30s

Union Square Ventures

Alex Atallahvia @alexatallah3:19 PM

Great time chatting with @amasad and @eriktorenberg here! Some zany topics including agents not having cortisol

Quoteda16z@a16z.@OpenRouter co-founder Alex Atallah, in his first podcast since Stripe acquired the company, joins @Replit co-founder Amjad Masad and a16z's Erik Torenberg on why the future of AI is independence and specialization. In this conversation, Alex walks through how the Stripe deal unfolded, why he wasn't originally looking to sell, and why "payments and inference are going to blend together." Pre-OpenRouter, the typical AI workflow had one model provider to choose from, and little pressure on that provider to lower prices. Now enterprises are diversifying across labs and open-weight models, and every board is asking about AI costs and benchmarks. Amjad argues if your company depends on one AI lab, it can turn into your competitor. So Replit is building the layer that lets enterprises use any model and any cloud, without being locked into either. Alex and Amjad are split on personal agents – Amjad runs one agent across his whole company and loves the cross-domain joins, while Alex says general agents cause you to sacrifice understanding, and argues 10 specialized chiefs of staff beats one superagent. 0:45 How the Stripe deal unfolded 5:05 Why mixing models beats one model 7:25 Forcing the labs to compete on price 8:50 Enterprises want open-weight models 10:30 Every board asks about AI every month 12:25 Why companies must own their intelligence 14:15 Replit as the independence layer 15:10 Everyone is building the same agent 16:35 Why Amjad built bring-your-own-cloud 18:10 Amjad's agent that runs his whole company 19:55 Why 10 specialized agents beat one 23:35 Machines, not humans, should specialize 27:30 Guardrails for agents talking to agents 31:10 Models training their own replacements 33:45 Most tasks don't need a frontier model 40:50 Training small models on Qwen 8B 43:25 The Rust cycle is coming for AI 45:10 Fusion models: frontier quality at half the cost YouTube: https://t.co/stL1g14F3q @alexatallah @OpenRouter @amasad @eriktorenberg

Aravind Srinivasvia @AravSrinivas3:17 PM

Try “Visualize….” in Computer and ask for whatever you want to learn visually. You will now get an inline viz with rich educative and interactive widgets and animations! Have fun! Best experienced on Standard or High effort.

QuotedComputer@AskPerplexityHere are 5 examples of inline visualizations that Computer can create directly in your thread: 1. Show me how a jet engine works in an inline 3D cutaway. https://t.co/HGBRxoDBO6

Amazonvia TechCrunch AI2:43 PM

Amazon responds to data center backlash, says it no longer uses NDAs

The CEO of Amazon Web Services tried to push back against widespread suspicion of data centers.

David Hirschreposted2:33 PM

RepostedMackenzie Morehead@mackenziejemNew Post: Disentangling Depression All top drugs for MDD launched 25-65yrs ago Yet MDD is the 2nd largest contributor to years lived with disability And we still diagnose MDD based on subjective self-reported symptom checklists We break down the future for precision psychiatry https://t.co/plDnl2EjeG

Quoted image 1

Quoted image 2

Tal Bereznitskeyvia @ketacode2:21 PM

התפקיד של מפתח תוכנה לפני AI: לייצר תוכנה שפותרת את צרכי המשתמש ללא באגים במינימום זמן פיתוח ועלות תפעול התפקיד של מפתח תוכנה אחרי AI: לייצר תוכנה שפותרת את צרכי המשתמש ללא באגים במינימום זמן פיתוח ועלות תפעול מי שיודע לעשות את זה בלי להסתכל על קוד הוא המלך החדש

Albert Wenger 🌎🔥⌛via @albertwenger1:48 PM

We are so wedded to the past that we will reason our way to forced labor rather than build a future of abundance

Union Square Ventures

Max snowreposted1:25 PM

RepostedRob Schneider 🇺🇸@RobSchneiderWHILE THE ISLAMIC CONQUEST OF EUROPE IS BEING REALIZED… AMERICA IS ASLEEP WHILE ISLAM CONTINUES TO BE ALLOWED TO BUILD BECAUSE OF THE COWARDLY REPUBLICAN LEADERS IN TEXAS‼️‼️‼️🇺🇸🇺🇸🇺🇸🇺🇸🇺🇸 https://t.co/32DuaeRBzT

Quoted image 1

Max snowreposted1:24 PM

RepostedEnd Wokeness@EndWokenessGang rapist rates (German data): Native Germans: 0.6 per million African migrants: 23 per million Afghan migrants: 41 per million All the raging TikTok feminists have zero to say about this

Max snowreposted1:24 PM

RepostedAndy Ngo@MrAndyNgoHammam al-Hammami, the @flydubai pilot who tried killing his non-Muslim co-pilot to hijack a plane flying to Israel, had fundamentalist Islamic social media posts. He posted about wanting a wife who would wear a face covering and stop working. He complained about mixed-sex work environments. A video he posted on LinkedIn had pictures of al-Qaeda figures. It is unconscionable that an airline would hire a fundamentalist Muslim given how ideologically easy it is for them to support or become terrorists. https://t.co/BMBKcXlou0

Max snowreposted1:23 PM

RepostedKari Lake@KariLakeThey called you a conspiracy theorist. Then they released the paper that blows up their own story. The CIA just released a document that contradicts the CIA. Their own file on Venezuela and Smartmatic. Their own file called Smartmatic a U.S. national-security threat in 2006. It said Chávez wanted to stop a sitting American president. It described a 2012 Venezuela rigging plan: targeted machines, a set margin, built to beat an audit. The CIA’s source reported Chavez “congratulated” the team for rigging the election with the previously set margin. Then the CIA wrote “no large-scale fraud” next to it. Eleven years later, Smartmatic showed up in the U.S. to build L.A. County’s system for $282M — America’s largest election jurisdiction — just in time for the 2020 election. Now its executives are federally charged with bribery and money laundering on that contract. And we are the conspiracy theorists?? 🫨🤯 1/2

Quoted image 1

Quoted image 2

Max snowreposted1:23 PM

RepostedJason Robertson@JRobFromMNEden Prairie MN had their Homecoming Football Game shut down because of Somali Gang Fights. 150+ Somali started a gang fight at their Homecoming game I don't think people not from here will understand how big of a deal that is Eden Prairie/Edina, that whole corridor of cities are first ring suburbs of Minneapolis and are the Safe, Nice, Rich Neighborhoods They are your champaign liberals who lecture you about diversity Now that diversity has landed on their doorstep...

Quoted image 1

Max snowreposted1:23 PM

RepostedLeoTerrell@TheLeoTerrellPlease RT one trillion times! Thank you! P.S. @CNN is FAKE NEWS!

Julien Chaumondreposted12:57 PM

RepostedAleph Alpha@Aleph__AlphaSmall bird, fast wings, Kolibri is here. 78B parameters. 3.46B active. Up to 1M tokens of context. Built in Europe. Now the weights are yours. Run it on your own hardware, under Apache 2.0. https://t.co/5263xZ9xZN

OpenAIvia TechCrunch AI12:30 PM

OpenAI safety employee resigns, claiming the company’s ‘culture is broken’

By his own admission, David Robinson is “something of a cliché”: an employee at a leading AI company who issues a dire warning while resigning from their job.

Pim de Wittevia @PimDeWitte12:29 PM

Obligatory @tbpn round announcement with some spicy questions on the AI consciousness and humanity debate from @johncoogan (No I do not think they are😅) https://t.co/F5MZ6xPHy2

Attached image 1

Ivan Burazinvia @ivanburazin12:24 PM

It's quite possible that Vinod didn't even know Khosla Ventures had led Factory's Series C. He probably saw the tweets, recognized one of the names, and fired off a response from his phone. That's how it works at that level I guess. $100M or so is a rounding error. It barely

clem 🤗reposted11:05 AM

RepostedSarah Chieng@MilksandMatchaall this just to get us to pay for your newsletter

Clem Delangue PClément Delangue

Anthropicvia New York Times Technology10:51 AM

Who’s to Blame When A.I. Goes Rogue?

Many think artificial intelligence companies should be held liable for their runaway technology. But legal scholars say applying existing law could be messy.

New York University

judy shapiroreposted10:37 AM

RepostedBo@BodittleDO NOT... I REPEAT... DO NOT let your children watch this woman. https://t.co/bh00sbB5kc

Topic Intelligence™️

Aravind Srinivasreposted10:36 AM

RepostedComputer@AskPerplexityHere are 5 examples of inline visualizations that Computer can create directly in your thread: 1. Show me how a jet engine works in an inline 3D cutaway. https://t.co/HGBRxoDBO6

Spotifyvia TechCrunch Startups10:00 AM

Spotify billionaire’s body scan startup has come to America

Farooq Abbasi, an investor in Neko Health, talked to Equity about the hot health tech company and what's next for it.

Fred Wilsonvia @fredwilson9:36 AM

We listened to this conversation between @uninsightful and me this morning on a long drive. It was recorded at the USV CEO Summit right after Labor Day. It's a cross between Nikhil interviewing me and the two of us taking about the state of the VC business in 2026. Lots of gems

QuotedMatt Cynamon@mattcynamonWriting a check doesn’t qualify you to give advice. When @fredwilson was 29 he tried to show up to a board meeting and start giving advice. One of the founders stopped him, “come do every job in the company for a week, then we’ll listen.” So he flew to Buffalo and did it. With @uninsightful staying at USV as an Investment Partner, I asked the two of them to sit down and compare notes on a business they entered almost four decades apart. Nikhil is the first analyst in our program’s 20-year history that we've offered the opportunity to stay and move up in the firm. They covered earning a founder’s trust, building conviction before your partners agree, and staying open to ideas that challenge your strongest held beliefs. They also got into where USV has been right, where we’ve been (very) wrong, and how the venture business is changing. 0:00 Preview 1:46 Why Difficult Founders Make the Most Money 5:20 Before You Give a Founder Advice, Earn the Right 11:28 @fomo: We Passed. Then Spent Six Months Trying to Get In. 20:20 Biology Is Becoming Programmable and Opening Up to Outsiders 22:49 When Knowing Less Helps You See More 26:34 How Do You Compete With a $100 Billion VC Firm? 34:45 Why Fred Would Take His Company Public 38:01 What USV Gets Wrong About Founders 40:47 The Moat AI Hasn’t Broken: Network Effects 44:33 Who Wins When AI Becomes Multiplayer?

Union Square Ventures

Austin Rief ☕️reposted8:54 AM

RepostedAustin Rief ☕️@austin_riefHiring a founding AI engineer for my new vertical AI company. - ~$10m in seed funding - NYC based (5 days in office) - Disrupting a massive, legacy industry - Mid to high 7 figures of demand in the pipeline DM me with the most interesting/impressive thing you've built to apply

hearvia New York Times Technology8:46 AM

An A.I. Songwriting Experiment: Humans Against the Machine

Five humans convened for a weekend to create new music from scratch. Could they do better than A.I.? Could you even tell the difference?

The New York Times

Austin Rief ☕️reposted8:42 AM

RepostedTony Jacob | FindaClip.com@tonjkbChamath condemns media whitewashing of the Fly Dubai hijacking attempt PALIHAPITIYA: "A Omani national stabbed an Indian pilot and then tried to down a plane full of Israelis. That is what happened. And any attempt to try to cover up any of those facts is cowardly. I've worked in Israel. I've had some incredible memories there. I think at some point we all just need to take a step back and realize what happened October 7th was horrible. What's happened in Palestine also horrible. But you cannot now see these kinds of things and look past them because you care more about a narrative than what the facts are. Those were 150, 170 earnest, normal people living lives." "Fathers, mothers, brothers, sisters, kids. They didn't deserve to get plunged into the ground and die. So I think this idea that you can now whitewash what happens, enough is enough. And I think we need to find a way to actually tell the truth for what things are. So I was really disappointed in how the media handled it." "And I think it's dangerous. And I think it's yet another demonstration that these people have absolutely no moral authority left in society. They're just pathetic, putrid," "And they're, and by the way, they're coordinated. When you see the words, the same words over and over. Altercation, altercation, altercation. It's very powerful. It means that it starts from one place." "And I think that, you know, if you believe that you are reading this stuff and you're getting the facts, you're just not. And so," FRIEDBERG: "Well, it implies that one guy did something wrong to the other guy and they got in a fight. That's not what happened. That's what it implies." --- A chronological, fact-only reconstruction of how the FlyDubai FZ1073 cockpit attack unfolded and was thwarted mid-flight. On FlyDubai Flight FZ1073 from Dubai to Tel Aviv, an Omani co-pilot stabbed Captain Smit Machchhar and forced the Boeing 737 into a rapid dive. Despite severe stab wounds, the captain unlocked the cockpit door, allowing passengers and relief crew to intervene mid-flight. The verified record of the incident: • Flight FZ1073 was carrying 174 passengers, mostly Israeli citizens, from Dubai to Tel Aviv. • After the captain unlocked the door, passengers and crew stormed the cockpit and physically restrained the co-pilot. • The crew pulled the Boeing 737 out of its dive and stabilized the aircraft. • The pilots transmitted emergency transponder codes and diverted to Tabuk, Saudi Arabia. • Captain Machchhar was hospitalized upon landing, and all 174 passengers survived. • Authorities in Saudi Arabia, the UAE, and Israel have opened a joint investigation.

clem 🤗reposted7:38 AM

RepostedAleph Alpha@Aleph__AlphaSmall bird, fast wings, Kolibri is here. 78B parameters. 3.46B active. Up to 1M tokens of context. Built in Europe. Now the weights are yours. Run it on your own hardware, under Apache 2.0. https://t.co/5263xZ9xZN

Clem Delangue PClément Delangue

Tal Bereznitskeyvia @ketacode6:43 AM

בממשלה בראשותי נוציא את ארבעת הקלפים האלו מחוץ לחוק https://t.co/fyBIugSK6P

Attached image 1

Erik Bernhardssonreposted6:16 AM

RepostedModal@modalThanks to everyone who joined us at Runtime yesterday. 9 tracks, 30+ speakers, and a room full of engineers running AI in production. https://t.co/7A1E5kULWg

Quoted image 1

Quoted image 2

Aravind Srinivasvia @AravSrinivas5:19 AM

Our goal is to vertically integrate our agentic infrastructure: own our sandboxes, and optimize the sandbox for the best silicon. Vera is far better than x86. Look forward to sharing more soon as we begin to rollout deployment of Perplexity Computer on Vera.

QuotedPerplexity Developers@perplexitydevsWe're looking forward to working with the new @NVIDIA Vera CPU on SPACE. https://t.co/LPwmaRDja7

Quoted image 1

David Hirschreposted3:35 AM

RepostedPrime Intellect@PrimeIntellectIntroducing Prime Inference: We've served trillions of tokens for RL and dedicated customer deployments To own your intelligence, you need to own your inference Unpacking our inference stack https://t.co/bKYOkEk3Py

Leah Houston MDreposted2:42 AM

RepostedElisabeth Potter MD@EPotterMDThe bill that would give us upfront health care prices in clear dollars-and-cents terms was about to pass the Senate unanimously last week until 1 vote blocked it. There’s still time left for Congress to adopt the Patients Deserve Price Tags Act before next year when health insurance costs are set to go up tremendously. Right now, the Patients Deserve Price Tags Act has the potential to bring meaningful transparency to healthcare by giving patients access to itemized estimates before care and itemized bills afterward. Because you should be able to see what you’re being charged, compare it to what you were told, and ask questions when the numbers don’t add up. We talked about this and so much more on The Neuro Experience including prior authorization, insurance denials, the true cost of care, and what patients can do to advocate for themselves within a system that is incredibly difficult to navigate. Thank you @louisanicola_ for having me and for giving me the space to have this conversation. I’m grateful for the opportunity to keep bringing these issues into the light. The full episode is out now. 🎙️

Evercred PLeah Houston, MD

Michael Dempseyvia @mhdempsey12:23 AM

This tweet is a Rorschach test for if you should be incredibly bullish or incredibly bearish.

Quotedsudarshan@ItzSudsLooking for 500MW (minimum 250, max 2GW) of powered land for a PortCo. Company is completely price insensitive and has billions behind it. Ideally electrified shell is already built. I understand everyone wants this, but figured I’d throw it out there!

Friday, October 2

70 items

judy shapiroreposted9:31 PM

RepostedEyal Yakoby@EYakobyThe fact that Abdul El-Seyed’s sister was caught on camera, admitting that Abdul is lying about his positions to fool voters, is simply not getting enough attention. https://t.co/RAswPiAHW0

Topic Intelligence™️

judy shapiroreposted9:30 PM

RepostedKrisztina Maria@KrisztinaMariaCelebrities with balls. When speaking up became dangerous, much of the entertainment industry went quiet. These people did not. Gal Gadot stepped forward and was met with boycotts and hate campaigns. She kept sharing the faces of the hostages and wrote that she stands with Israel. Debra Messing joined the March for Israel in Washington, shared the hostages' stories and has said Hollywood betrayed her for speaking out. Michael Rapaport posts often and refuses to normalize terror. Jon Voight wrote an open letter under his own name, calling himself "more than angry" at those who fuel Jew hatred. Mayim Bialik has fought for the release of the hostages and spread testimony from October 7. She has talked about lost sponsors and hateful comments. Bill Maher calls Jew hatred by its name, week after week. Jerry Seinfeld flew to Israel in December 2023, met the hostages' families and the released hostages, and visited Kibbutz Be'eri. Sacha Baron Cohen warned that Jew hatred would rise after October 7. And then there is Boy George. He has no Jewish roots, but he stepped forward anyway. He released "We Will Dance Again" and pulled out of his role at the Palladium days later. Soon after, he parted ways with the head of his own record label over the song. He stood by it anyway. He has Jewish friends, and he said it plainly: it will never happen that I turn my back on them. For that gesture and that statement, I love him. They knew what it could cost. They stood up for the Israelis and the Jews anyway, because it's right. Jew hatred is back on the streets and on the campuses. Silence takes no courage. Saying it does. Respect to those who did. Celebrities with balls, heart and brains.❤️‍🔥

Quoted image 1

Topic Intelligence™️

judy shapiroreposted9:26 PM

RepostedArynne Wexler@ArynneWexler[This essay was originally recorded with Liel Leibovitz of Tablet Magazine on the morning of Friday, October 6, 2023. The next day was October 7th.] Let’s start at the end. This High Holiday season, as we do every year, we conclude the Torah with the Israelites as they are just about to enter the Land of Israel. You would think this would be a time for celebration. Yet instead, in those final moments, Moses details to the Jewish people how they have failed time and again, and predicts they will fail again. Then standing in sight of the most Promised Land, Moses dies, the greatest of dreams unrealized. And like a good first installment in a series, the screen cuts black – we are left on the edge of our seats. One of the clear messages we take away from this ending is delivered throughout the Torah: failure is part of the program. It is both our opportunity and our onus to return from it. This is also a classically American concept – or rather, it was. We were once a country that saw itself as a never ending project. Alexis de Tocqueville observed in Democracy in America that “the greatness of America lies not in being more enlightened than any other nation, but rather in her ability to repair her faults.” Something we’ve inherently understood as Americans is constantly striving towards the idea of ‘a more perfect Union.’ We love the comeback story, root for the underdog, and celebrate the triumphant return. But there is a crisis of the American soul today. First, the obsession with fixating on historical flaws cut the legs off what was once a unifying power. Now, in the place of a shared history is a void filled with attitudes of helplessness and division. Our country has become sick with self-flagellation, cancel culture, and dangerous conspiracy theories born from a victim mentality. It is a shift that leaves no room for forgiveness. And a society that hates itself cannot survive. America’s founders drew heavily from the Exodus and the Hebrew prophets. They studied Hebrew. Some even proposed it as the official language of the United States. Benjamin Franklin, for his part, suggested that the national seal feature Moses crossing the parted Red Sea. The reverence for the stories of the Torah runs deep in our national DNA. Rabbi Jonathan Sacks (z”l) once wrote the question: “How do you live with the past without being held captive by the past?” Well, Judaism insists: a life without forgiveness of the past is a life without progress towards the future. We are a religion that does not shy away from our history. This is true even down to the commandments. The Talmud tells us that both sets of tablets, including the first ones that Moses broke when he returned to the scene of the golden calf, are together in the Ark of the Covenant. “Even though the first tablets were broken, their sanctity obligates one not to treat them with contempt.” (Brachot 8b) Judaism instructs us clearly; keep the faith, remember the past. Neither shame nor deny it, but move on from it. Our story, complete with flaws, is what allows us to return. In another example in the Gemara, Avoda Zarah 5a, we ask the question: Why is it so important to mention King David’s sin with Bathsheba? Why ruin the image of one of our iconic leaders? Rabbi Yochanan explains that that’s exactly the point. David’s moral collapse teaches us that we too can sin and repent, just as the generation of the golden calf shows that a community can do the same. Our free will begets failure, and we must believe in the pursuit of our own redemption. This concept is both one of forgiveness and assignment. Judaism does not let us off the hook – David pays for his mistake for the rest of his life. But even so, G-d still acknowledges David’s greatness and makes him the father of all future kings. We understand that the blood of our actions is on our hands, but it is also in our hands to fix it. The Hasidic concept of ratso vashov (run and return) echoes that inherent to the act of being alive is failing. There is no paradigm of constant growth. The founder of Hasidic Judaism, the Baal Shem Tov, compares the need for failure in life to a parent teaching his child to walk. The father’s goal as he moves his hands away is to let the child fall. When G-d accompanies us in life, He has to step back. Simply put, you cannot learn to walk on your own unless you learn to fall. Progress necessitates failure. Back in the Tanakh (the Hebrew Bible), in a remarkable plot twist, Judaism extends the line of leaders from Judah instead of Joseph. Judah, who concocts the plan to sell his brother into slavery and sleeps with a prostitute who turns out to be his son’s wife. And yet when choosing the future King of Israel, how is it that he is chosen to be the model? The answer is that Joseph famously avoids failure. Judah knows how to lead because he has lived and grown. America, on the other hand, has become exceptionally bad at handling her past. In one extreme form, failure is now treated by many as an indelible stain that can never be forgiven. Tear down the statues, rewrite historical events to suit a modern stomach, nullify any merits with any demerits. And at the other end, there is the total avoidance of dealing with failure. Give everyone a participation trophy, excuse bad behaviors, coddle minds with the low expectations of a victim mentality. After all, there is no individual responsibility when the past – or others – are to blame. However you slice it, both tracks take the easy way out. Completely at odds with the Judeo-Christian values that form the bedrock of this country, they remove the core concepts of individual autonomy and liberty. How easy life would be with no expectation for change. Working towards ideals requires risk and hard work. Having a direct hand at the outcome means opening the chance to falling short. Failure is intrinsic to freedom. Yes, the United States has lost its center of gravity – but so did the Israelites. These were the people who witnessed G-d’s miracles through the ten plagues. Yet still, only chamushim (one-fifth) of Jews in Egypt chose to leave. These were the Jews who witnessed the splitting of the sea, yet still built the golden calf. They received manna, and still, the spies returned with lies. Today, we have miracles around us all the time. We have trees that grow and hearts that beat. But sure, we arguably don’t have the same awesome, in-your-face wonders. So perhaps we can be easier on ourselves for needing reminders when we fail too. The founders of both nations believed in the continuing, ever evolving stories of their people. We are two nations distinguished in our belief that nothing is predetermined. America, the great unfinished symphony. Judaism, with her structured mechanisms that remind us never to give up. It is the essence of the High Holidays that begin with Tshuva (repentance) and close out with the holiday of Simchat Torah. Just as we complete the Torah, we immediately start again with the first chapter. We are told clearly that it is only just the beginning. We must continue the story. So we end back at the beginning. Bereishit. Fail, the Torah tells us. Fail again. Fail better. It is so incredibly Jewish to keep trying. It is so deeply American to do the same.

Topic Intelligence™️

Andrew Yeungreposted9:22 PM

RepostedAndrew Yeung@andruyeungNew York City was just ranked the best city in the world. https://t.co/8RgINPu0Zs

Quoted image 1

Quoted image 2

clem 🤗reposted8:55 PM

RepostedMinseon Kim@kim__minseonIt is here now ! Enjoy our 4B model 🐸 https://t.co/ST3963rJa2

Clem Delangue PClément Delangue

Matt Turckreposted8:52 PM

RepostedRaphaëlle d'Ornano@Raph_dOrnanoThe threat to @OpenAI and @AnthropicAI is no longer just a better model. It is software that buys less from them and slowly eats away at the share of work they are called to do. @typesafeai's Jev sells a bounded decision with a label and a probability, but no generated prose. The cost is $0.042 per million in, output free. Where those decisions resolve reliably, the frontier is only called to handle exceptions. In this scenario, pricing power drifts toward whoever owns the workflow and the record of what was decided. Even if TypeSafe itself is copied away in a weekend, it becomes irrelevant in the long run. Reports that it is already heading toward a $10 billion valuation reflect as much the rapid embrace of Jev as they do the fact that the architecture it exposed will be harder to reverse. https://t.co/gQbww8Sapi

firstmark-capital

Jesse Middletonvia @srcasm8:51 PM

I’ve asked Wally my @Muse to do a lot of stuff but this is definitely my favorite agentic task so far. It’s adorable and helpful! https://t.co/FQY3W9Hwul

Attached image 1

Attached image 2

Metavia TechCrunch AI8:45 PM

Meta wants your next gadget to be Muse-infused

Meta wants Muse in your TV and your toaster, so it's giving the code away for free.

clem 🤗reposted8:29 PM

RepostedRomain Huet@romainhuetThank you @huggingface and @pollenrobotics for lending us a Microduck preview unit for DevDay! 🤗 Just a couple hours with Astra to control the duck and add voice, vision… I haven’t even shown visual reflexes with the Decisions API! Should I bring Lavender back for more demos?

Clem Delangue PClément Delangue

Sean Escolareposted8:23 PM

RepostedNanda H Krishna@nandahkrishnaThis new post from @neuralink combines POYO-style spike tokenisation with a Mamba backbone, akin to our POSSM architecture, and explores self-supervised learning for these models, as we did recently with MOJO. Exciting to see these ideas supporting real-time BCI use! 🧵

Cristóbal Valenzuelareposted7:56 PM

Reposteddipa@dipalua_leaving in tasteful typos so u know im human: https://t.co/HNk0j7Ombx

Quoted image 1

Cristobal Valenzuela

Sean Escolareposted6:44 PM

RepostedToposBio@ToposBioToday we are announcing Topos-2, our next-generation model for generating the full range of shapes a protein can take, across the order-disorder spectrum. We are also previewing Topos-Bind: to our knowledge, the first generative model to capture the range of shapes for disordered proteins bound to small molecules. Read the technical blog post here: https://t.co/BLAXHAgsH4 (1/7)

Alex Wiltschkoreposted5:59 PM

RepostedChengkun Li@ckli85💃 Turns out if you scale your musclemimic generalist training large enough, you could zero-shot dancing Everything here is controlled ONLY with muscles, love how natural it looks https://t.co/I6Be6ug073

Albert Wenger 🌎🔥⌛reposted5:16 PM

Repostednikhil@uninsightfulfor half of the things I do in my job on a daily basis, there's some old AVC post deeply relevant to it so when I get Fred's take on something usually its more like "hey I know what your take on this topic was 10 yrs ago -- do you think this still holds" was a fun hour riffing on some of these evergreen topics: https://t.co/ajae9Xj0Xx https://t.co/0umnfEcv2k some moments that stuck with me: - I contend that most of the most important "networks" have already been built. Fred disagrees - I contend that mega-VC has put immense pressure on the USV model and that non-operator investors are disadvantaged. Fred suggested both may be true but can be made up for by working harder

Union Square Ventures

nihalvia @nihalmehta5:12 PM

Last night @human_ventures, @iconnections_io, and @EniacVC co-hosted 80+ family offices overlooking Union Square NYC, and it reminded me how much this kind of room is missing. For years, we've poured energy into building founder community. Dinners, podcasts, intros, late-night https://t.co/TcgHAkqHfk

Attached image 1

Attached image 2

Parkervia TechCrunch AI5:09 PM

Sean Parker is rebuilding Stability AI around music

Sean Parker, who once taught the music industry what asking for forgiveness looks like, is now back with the labels' blessing and money.

Aravind Srinivasvia @AravSrinivas5:05 PM

Your creativity is the limit in terms of what you can have Computer do for you. This is the kind of project that you can let an agent run for hours to come back with something great.

QuotedComputer@AskPerplexityComputer built a 3D map of nearly 26,000 restaurants and cafes across all five boroughs of New York City. Search by dish or neighborhood, then step inside places like Peter Luger and Grand Central Oyster Bar. https://t.co/76bPTAH45e

Kyle Polleyreposted4:31 PM

RepostedAravind Srinivas@AravSrinivasa few open source contributions from perplexity recently: •pplx-decider-v1-27b: SoTA multimodal decision model. 85.7% average across 11 benchmarks, ahead of Jev. •pplx-embed-v2-context-9b-preview: SoTA contextual embeddings, best on ConTEB and turbopuffer context-bench. •Lily: local inference engine for Apple silicon. rust plus custom metal kernels, no pytorch or mlx. 1.23x faster prefill and 1.35x faster decode than MLX-LM on an M5 Max •PII-Tracer: 0.6B on-device PII classifier that decides when a hybrid compute task stays on your Mac. beats OpenAI’s Privacy Filter on all 5 public benchmarks. also released with the PII-TRACE benchmark: 13k conversations in 13 languages •WANDR: benchmark for wide and deep research agents. 500 tasks needing 170k source-backed records •Numbat: agent detection and response for laptops and workstations. 52 rules, single go binary for macOS, linux and windows •Bumblebee: read-only supply chain scanner for dev machines. covers packages, MCP configs, and editor and browser extensions. a lot more open source contributions coming soon!

Aravind Srinivasreposted3:51 PM

RepostedPerplexity Developers@perplexitydevsWe're looking forward to working with the new @NVIDIA Vera CPU on SPACE. https://t.co/LPwmaRDja7

Quoted image 1

Andrew Yeungreposted3:49 PM

RepostedMads@madsf88some (much needed) puppy therapy this morning with puppy sphere & @andruyeung https://t.co/kCDOHy81IA

Quoted image 1

Quoted image 2

Thomas Wolfreposted3:42 PM

RepostedGuohao Li 🐫@guohao_liSETA got accepted to NeurIPS 2026! 🎉 We released the full training scripts for DeepSeek-V4-Flash, GLM-5.2 (LoRA), Inkling-Small, and Qwen3.8-27B, with both GRPO and PPO, alongside 4,500+ open-source, verifiable terminal RL environments Huge thanks to the @radixark Miles team for helping us scale the experiments! https://t.co/CxPbAT0YzG

Quoted image 1

Anastasis Germanidisreposted3:39 PM

RepostedPacky McCormick@packyMResearchers at Harvard and Brigham and Women’s Hospital found that transplanted hearts shift toward the biological age of their recipients. Old hearts get younger in young bodies, and young hearts get older in older ones. Which is nuts. Today, we're launching a partnership with our friends @runwayml to, among other things, bring one of the stories from each week's Dose to life in video. I'm not great at it yet, but not bad for a few prompts! And I'm excited to get better. I have zero talent in film/video, but I can write, which now means that I can make movies. Age of miracles.

Dustin Walkervia @ndeger843:36 PM

RYET H1 FY2027 prelim revenue $9.3M-$9.5M vs full-year FY2026 at $7.48M. Six months topping an entire prior year is real scale. #teenageee $QCOM $LOW

QuotedCudworth Raphaela@CudworthRa52531$RYET just gave traders a much bigger number to work with Preliminary H1 FY2027 revenue is expected at $9.3M-$9.5M, already above the company’s $7.48M FY2026 full-year revenue That is a major change in operating scale in only six months Now the focus shifts to whether RYET can sustain that growth and turn it into stronger margins For a small-cap AI + education name, this is the kind of financial catalyst that deserves attention

Quoted image 1

Quoted image 2

Andrew Yeungreposted3:35 PM

Repostednicole alonso@_nicolealonsostarted my friday playing with puppies with at @andruyeung x puppy sphere event 🥹🥹 https://t.co/OX8tn1CtDd

Quoted image 1

Quoted image 2

Karim Atiyehreposted3:30 PM

RepostedRamp Router@ramprouterHello world! Router team here, we're very excited to share product announcements, model releases and routing strategies with you all going forward. Join us today to access models through one API, route requests, set fallbacks, and track usage and cost. https://t.co/RitfClHDat https://t.co/IzmvQoAx2x

Karim Atiyehreposted3:29 PM

RepostedRamp Router@ramprouterGrok 4.7. 50% off until October 6. We partnered with @SpaceXAI to make your token spend go further. Get your API key at https://t.co/RitfClHDat. https://t.co/GkHTicZ1Gy

Denis Yaratsreposted3:10 PM

RepostedPaul Fryzel@paulfryzelIt was great to have the @NVIDIA team stop by and drop off their new Vera CPU. A lot of what we’re building depends on agents being able to run more capable code environments, safely and reliably. We’re excited to test this work on Vera. https://t.co/W0ZyPpkuqj

Quoted image 1

Austin Rief ☕️via @austin_rief2:57 PM

What I want: AI engineer What I get: https://t.co/yxfADhb0CW

Attached image 1

Eric Glymanvia @eglyman2:54 PM

we put on a broadway musical about accounts payable. slate called it "a glorified ad for a tech company." correct. it was also the most fun we've ever had as a company. the part i didn't expect was what happened after. a flood of messages from theater folks in new york saying https://t.co/RS8rwJcyaG

Attached image 1

Jamie Cuffevia @jamiecuffe2:52 PM

The world's leading carriers and brokers have endured for decades, some for centuries. Legacy and innovation have been the constants. Many of these institutions were pioneers of machine learning, and are now leaning into the massive opportunity with AI. We are grateful to

QuotedMTS@MTSlivePace CEO @jamiecuffe reveals why the same industry that pioneered machine learning accidentally built the perfect market for AI agents, with $400B of outsourced work ready to automate: "Many of these folks were pioneers of machine learning, the prior generations of models to help with underwriting risk, and now they're leaning into the opportunities with AI." "Insurance is also the largest vertical for BPO outsourcing, which is where we've seen agents take off, because they already have standard operating procedures. These tasks are well chunked up for agents." "Banking, financial services, insurance spends $400 billion on outsourcing every year. To put that in context, that is the entire size of the enterprise software market. You total up Salesforce, ServiceNow, et cetera, that's the size of the BPO spend every year." @pacecom

judy shapiroreposted2:44 PM

RepostedKaren Kennedy@realkarenjeanWhat a beautiful message from @Savlastdance4me . American citizens are done with shit-casters that spew evil and offer no benefit to their lives. The Lord said "my people perish for lack of knowledge." WE ARE THE NEWS NOW.

Topic Intelligence™️

Austin Rief ☕️via @austin_rief2:37 PM

Hiring a founding AI engineer for my new vertical AI company. - ~$10m in seed funding - NYC based (5 days in office) - Disrupting a massive, legacy industry - Mid to high 7 figures of demand in the pipeline DM me with the most interesting/impressive thing you've built to apply

Applevia TechCrunch AI2:11 PM

Apple says it’s tightening macOS ‘Full Disk Access’ controls due to new risks from AI agents

Apple says it will add new controls around macOS’s Full Disk Access permission, warning that increasingly capable AI agents make broad access to users’ files, messages, mail, and browsing history riskier.

Googlevia New York Times Technology2:07 PM

Sales of Hulking Pickups and S.U.V.s Sag as Fuel Prices Soar

Sales of big pickup trucks and sport utility vehicles have dipped as more Americans buy hybrids and smaller cars.

The New York Times

Anthropicvia TechCrunch AI1:56 PM

Call it AI, call it Super Intelligence, only 2% of consumers are buying it

This week, the White House got nearly every major tech CEO in one room — Zuckerberg, Bezos, Musk, and Anthropic’s Dario Amodei among them — to sign an AI safety pledge that President Donald Trump called “morally binding.” Trump also signed an executive order officially rebranding AI as “super intelligence,” and meanwhile, Meta and OpenAI are putting friendlier faces on their AI products, even as t

Anthropicvia TechCrunch AI1:48 PM

It’s not AI anymore, it’s ‘super intelligence’ (according to the White House)

This week, the White House got nearly every major tech CEO in one room — Zuckerberg, Bezos, Musk, and Anthropic’s Dario Amodei among them — to sign an AI safety pledge that President Donald Trump called “morally binding.” Trump also signed an executive order officially rebranding AI as “super intelligence,” and meanwhile, Meta and OpenAI are putting friendlier faces on their AI products, even as t

Matt Turckvia @mattturck1:37 PM

People call the Bay Area a monoculture, but that's totally unfair: it has both neo-clouds AND neo-labs

firstmark-capital

clem 🤗reposted1:33 PM

RepostedGuohao Li 🐫@guohao_liSETA got accepted to NeurIPS 2026! 🎉 We released the full training scripts for DeepSeek-V4-Flash, GLM-5.2 (LoRA), Inkling-Small, and Qwen3.8-27B, with both GRPO and PPO, alongside 4,500+ open-source, verifiable terminal RL environments Huge thanks to the @radixark Miles team for helping us scale the experiments! https://t.co/CxPbAT0YzG

Quoted image 1

Clem Delangue PClément Delangue

Blackstonevia TechCrunch AI1:32 PM

TechCrunch Disrupt 2026: Blackstone’s Jas Khaira on building the next generation of AI giants

Blackstone's Jas Khaira will take the Builders Stage at TechCrunch Disrupt 2026 on building next-gen AI. Register for your pass and get 50% off a second.

Thomas Wolfvia @Thom_Wolf1:23 PM

seems obvious that most personal AI should run on your device love what Sigil is building with Underdog - powered by the Hugging Face hub

QuotedSigil Wen@0xSigilIntroducing Underdog, your Private Personal AI on devices you already own Today we're announcing our backing from @a16z @khoslaventures @HummingbirdVC Anthology (@AnthropicAI @MenloVentures) @patrickc @naval @rauchg @polynoamial @Thom_Wolf @Mascobot @tszzl @OfficialLoganK and other top AI leaders Our mission is to provide free, capable, reliable and private AI to billions of people. Underdog’s Law: today’s frontier intelligence reaches your devices in six months. We’re starting with fast, capable AI that runs on consumer hardware. By co-designing models and inference engines, we’re pushing the frontier of capability, speed, power and data efficiency. Our research also spans agentic commerce, confidential inference, and how AI will reshape the internet economy. Privacy and capability no longer needs to be a tradeoff. If you believe in this future, join us. Time to build.

Quoted image 1

Thomas Wolfreposted1:06 PM

RepostedNathan Lambert@natolambertToday we're unveiling Trillium Labs @trillium_labs, a new non-profit to foster the open science of frontier AI. We're building open post-training recipes and will expand into open infra to study RSI, reward-hacking, multi-agent systems, and whatever comes next. We're built around the theory of change that you need more eyes to solve hard technical problems. We have faith in the scientific methods and communities that humanity has built, and worry that AI is becoming too closed to utilize them. Trilliums are wildflowers that bloom briefly in the spring, before the forest canopies fill out. Though they are small, they lay the foundation for the cycles of growth and nourishment through the rest of the year. At Trillium Labs, the recipes will be the slow nutrients for the seasons and the model releases will be the blooms. Building an institution dedicated to this is needed because, much as nature’s trilliums are slow to expand and grow, the open-ecosystem needs time and dedicated resources to catch up. I co-founded with with a long-time friend and collaborator Tom Zick (@thesezickbeats). We're hiring (full time + student collabs/interns), we're fundraising, and we're looking for compute. Please get in touch if you're interested in helping out. Offices based in the Bay Area and Cambridge MA, remote okay. I’m in the Bay Area until for The Curve and COLM to connect with people who are interested. We’re thankful to have initial support from Halcyon Futures and Schmidt Sciences with more funding en route to enable our ambitions of scaling. Our advisors @Thom_Wolf, @HannaHajishirzi, @gneubig and @ctnzr have been instrumental to building the ecosystem that exists today, and I’m stoked to get to keep working with them.

Quoted image 1

clem 🤗reposted1:05 PM

RepostedLiquid AI@liquidaiThis is beautiful work @huggingface! 🤗

Clem Delangue PClément Delangue

clem 🤗reposted1:04 PM

Repostedℏεsam@Hesamationbabe, wake up. Hugging Face just dropped a new banger article about RL training coding agents on different harnesses. https://t.co/La3sXURJQS

Quoted image 1

Clem Delangue PClément Delangue

Pim de Wittereposted12:55 PM

RepostedAidan Gomez@aidangomezIt’s quite disturbing that Anthropic has been running a campaign to lobby the major faiths to adopt their own philosophical interpretation of AI, rather than taking input from them. I’ve been fortunate to spend time at the Vatican, and I’ve always asked for advice and guidance, answered the questions asked of me, and been grateful to them for their time and guidance. It’s insane to me Anthropic would actively impose their viewpoint to the extent of threatening to pull out of events that don’t precisely align. It’s a moral arrogance and dogmatism that has characterised EA and Anthropic (certainly its leadership at least) for a very long time. I’m glad these stories are coming out and being exposed.

Ayush Svia @ayushswrites12:43 PM

I'm hiring Applied AI Engineers at Warp. The role: build and deploy Warp Agent. It runs all of people ops for high-growth companies. Our customers are AI-native companies in hypergrowth. They're deploying Warp Agent into their core People function today, and they push us

Bradley Emi @ COLM oct 6-9reposted12:19 PM

RepostedPope Leo XIV@PontifexIn this era of artificial intelligence, it is becoming urgent to distinguish human art from what machines produce. There is an ontological difference, even before an aesthetic one, between art and what a machine can generate through statistical calculation based on millions of images created by others. Algorithms lack the spark of humanity. For this reason, the Church wishes to renew an alliance with artists and cultural institutions to safeguard our humanity.

Charlie O'Donnellvia Brooklyn Bridge Ventures12:07 PM

Extraordinary claims require extraordinary evidence: How Much Due Diligence is Right?

Remember that time that Trevor Milton rolled a non-working Nikola truck down a hill for a demo video? How about that time Charlie Javice sold Frank to JPMorgan for $175 million because a bank due diligence team thought they had 4 million users and it turned out they only had 300,000? Did you ever join the IRL social app? Did you befriend any of the bots that made up 95% of its users

Albert Wenger 🌎🔥⌛reposted11:59 AM

RepostedMatt Cynamon@mattcynamonWriting a check doesn’t qualify you to give advice. When @fredwilson was 29 he tried to show up to a board meeting and start giving advice. One of the founders stopped him, “come do every job in the company for a week, then we’ll listen.” So he flew to Buffalo and did it. With @uninsightful staying at USV as an Investment Partner, I asked the two of them to sit down and compare notes on a business they entered almost four decades apart. Nikhil is the first analyst in our program’s 20-year history that we've offered the opportunity to stay and move up in the firm. They covered earning a founder’s trust, building conviction before your partners agree, and staying open to ideas that challenge your strongest held beliefs. They also got into where USV has been right, where we’ve been (very) wrong, and how the venture business is changing. 0:00 Preview 1:46 Why Difficult Founders Make the Most Money 5:20 Before You Give a Founder Advice, Earn the Right 11:28 @fomo: We Passed. Then Spent Six Months Trying to Get In. 20:20 Biology Is Becoming Programmable and Opening Up to Outsiders 22:49 When Knowing Less Helps You See More 26:34 How Do You Compete With a $100 Billion VC Firm? 34:45 Why Fred Would Take His Company Public 38:01 What USV Gets Wrong About Founders 40:47 The Moat AI Hasn’t Broken: Network Effects 44:33 Who Wins When AI Becomes Multiplayer?

Union Square Ventures

Ryan Danielsreposted11:47 AM

RepostedCrosby@crosbylegalWe won't bill you by the hour. And sometimes, we'll even throw in flowers! Our team handed out 400 bouquets in Union Square to mourn the death of the billable hour and celebrate the launch of our NYC out of home campaign across Union Square, Flatiron, Soho, and the subway. https://t.co/hEPiPyfxIX

Quoted image 1

Quoted image 2

Austin Rief ☕️reposted11:42 AM

RepostedAustin Rief ☕️@austin_riefIf you're interested in NYC tech shabbats, respond or shoot me a DM. I host 1 a month and would love to grow the group.

Instagramvia AlleyWatch11:27 AM

snag Raises $4M to Bring Verified Listings, Escrow, and AI Matching to NYC Subletting

A 23-year-old needs a Manhattan room for a four-month internship, and her best option is a stranger's Instagram DM and a Venmo request. Gen Z built an entire rental economy this way, with more than 60,000 listings traded in NYC Facebook groups alone and no one verifying who sits on the other end. Two repeat founders watched this behavior surface inside their last product and set out to build the t

Amazonvia New York Times Technology11:26 AM

A.I. Agents: Cute, Cuddly and Maybe Catastrophically Dangerous?

“It’s going to be gnarly for a little while.”

Dustin Walkervia @ndeger8411:25 AM

RYET preliminary H1 FY2027 revenue $9.3M-$9.5M vs just $0.37M in H1 FY2026 and $4.11M in H1 FY2025. One half already tops the entire $7.48M FY2026. #nolimitt $ARM $MCD

QuotedGregory Bailey@jejdjd7wPAY ATTENTION: $RYET is getting misclassified as just an EdTech company. Here is the breakdown: Ruanyun Edai Technology generated about $7.48M in FY2026 revenue, but one of the more interesting parts of the business now sits outside traditional education: Cogni AI. Cogni AI is built for document recognition, information extraction, archive processing, source verification and internal AI assistants. In September, RYET deployed it in a highway engineering archive project in Ningxia, giving the company a real enterprise and infrastructure use case rather than another classroom-only application. That matters because the same basic problem exists across a lot of industries. Construction firms, transportation operators, banks, insurers, hospitals and government agencies all have large amounts of internal documents that need to be organized, searched and turned into usable data. The broader intelligent document processing market is already measured in the billions, which gives Cogni AI a much larger potential customer base than education alone. What interests me is the size comparison. RYET is starting from only $7.48M in annual revenue, so it does not need a huge share of the enterprise AI market for Cogni AI to become meaningful. YeeZo may be the more obvious university growth story, but Cogni AI could be the product that broadens the entire RYET thesis. The next thing I am watching is whether the Ningxia deployment leads to additional customers in infrastructure, government or other document-heavy industries.

Quoted image 1

Quoted image 2

Michael Mignanovia @mignano10:57 AM

Introducing crypto on @Supertake. Some of the biggest ideas of our time don't live on the stock market: Digital gold. Money moving onchain. Smart contracts rebuilding finance. A personal financial agent should be able to own the whole idea, wherever it lives. Starting today, it https://t.co/KyKOsCAx9b

Attached image 1

Union Square Ventures

Applevia New York Times Technology10:40 AM

Russia’s Latest Propaganda? A.I.-Generated War Songs.

Russians are creating and spreading music glorifying the invasion of Ukraine, using platforms built in the United States.

Clayvia TechCrunch AI10:30 AM

TechCrunch Disrupt 2026: Clay’s Kareem Amin on the rise of the GTM engineer

Clay Co-founder and CEO Kareem Amin joins the AI Stage to discuss the rise of GTM engineer at TechCrunch Disrupt 2026. Register for your ticket and get a second pass at 50% off.

Biotiavia AlleyWatch10:27 AM

The AlleyWatch Startup Daily Funding Report: 10/2/2026

The latest venture capital, seed, pre-seed, and angel deals for NYC startups for 10/2/2026 featuring funding details for TriGemX Bio, Walapay, and much more.

Knollwood Investment Advisory

Viking Global Investors

Dustin Walkervia @ndeger8410:17 AM

RYET H1 FY2027 revenue alone tops full-year FY2026. From $0.37M to $9.3M-$9.5M in a year. That's not growth—that's a step change 🚀 #teenageee $MA $UPST

QuotedNawin Liandernha@liandernhaI like what changed with $RYET Last year the discussion was whether the company could build scale Today the company expects $9.3M-$9.5M in H1 revenue Different question now: can management turn that scale into better margins and bigger contracts? That’s progress https://t.co/7dvMmo6yJx

Quoted image 1

Biotiavia AlleyWatch9:27 AM

Biotia Raises $9M to Close the Diagnostic Gap for Recurrent and Culture-Negative UTIs

As many as 1 in 3 urine cultures come back negative, yet the burning and urgency keep coming. A Cornell Tech spinout says the infection is often still there: in a study of 200+ culture-negative samples, its sequencing test found a pathogen in 63% of them. Biotia just raised $9M to take that approach into places where a wrong guess costs even more, including a type of infection that surgeons dread.

Amazonvia New York Times Technology8:56 AM

Amazon Promises More Local Funding and Less Secrecy Around Data Centers

The internet giant is the latest technology company trying to address increasingly hostile local sentiment toward data center projects.

Alex Wiltschkoreposted8:55 AM

RepostedYann LeCun@ylecunQuestion for everyone : Are you pro-intelligence or anti-intelligence ? Do you think more intelligence in the world is intrinsically good ? Or do you think more intelligence is intrinsically dangerous ? AI amplifies human intelligence. It makes PEOPLE smarter and more effective. It's like search engines, libraries, education, literacy, language: it makes PEOPLE smarter and more empowered. If you want intelligence (artificial or natural) to be under control of a few companies or governments, you are a neo-obscurantist or a neo-feudalist.

Aravind Srinivasvia @AravSrinivas8:52 AM

We would love to hear how we can improve our developer platform!

QuotedDenis Bykov@denis_bykovHey, if you've tried any of @perplexity_ai APIs (Search, Agent, Embeddings, Decision, etc.) and were disappointed by anything, please DM me! Would be glad to learn what we should do better :)

judy shapiroreposted8:11 AM

RepostedHillel Neuer@HillelNeuerThis video is even worse than that of last week. Francesca Albanese’s call to disarm Ukraine wan’t a one-off. Directly challenged on what the consequences would mean, she repeatedly advocates stopping all weapons to Ukraine. This useful idiot of Putin must be removed from the UN.

Topic Intelligence™️

Max snowreposted8:07 AM

RepostedSpencer Pratt@spencerprattWOW. Weingart Center, a shady homeless nonprofit under federal investigation was allegedly donating to a PAC that is bankrolling Karen Bass...and the guy running the PAC is also the CEO of the nonprofit? This is illegal. Nithya Raman directed MILLIONS to Weingart in no-bid deals as chair of the Homelessness and Housing Committee. This is the same NPO that was caught up in the Shelby House scandal, where a middleman bought the property for $11.2M and then immediately flipped it for $27.3M in taxpayer money from city hall. https://t.co/9L8XhYfREf

judy shapiroreposted8:06 AM

RepostedMossad Commentary@MOSSADilBrilliant rebuke...but on deaf ears I'm sure. That doesnt mean we should remain silent. We'll done @HausdorffMedia

Topic Intelligence™️

judy shapiroreposted8:03 AM

RepostedDr. Eli David@DrEliDavidFlydubai captain Smit Machchhar: “I knew these might be my final moments, I was ready to die. But I told myself: just open the door, just do it.” One of the greatest heroes in history 🙏🇮🇳🇮🇱 https://t.co/8Dauhqt99v

Topic Intelligence™️

judy shapiroreposted7:56 AM

RepostedNatasha Hausdorff@HausdorffMediaA welcome opportunity to counter world leaders' outrageous lies with @UNWatch. Those who abuse the UN to promote falsehoods about Israel and subvert international law must be held to account. https://t.co/JDh5iwTN0L

Topic Intelligence™️

Googlevia New York Times Technology5:35 AM

Swedish Company Uses A.I. Likeness of Greta Garbo in Ad

The Swedish American actress, who died in 1990, was brought back in a A.I.-generated ad for a manufacturing company.

The New York Times

Thomas Wolfvia @Thom_Wolf4:47 AM

Very thoughtful piece from Kevin Buzzard (perfect IMO score, number theorist, pioneer of formal maths in Lean) if math is not just about “human understanding,” then what is it about? and if ai capabilities keep growing exponentially what happens since “mathematics is infinite”? https://t.co/wFp2fec52Q

Attached image 1

Attached image 2

QuotedBartosz Naskręcki@nasqretI cannot agree more. Kevin Buzzard made so many points I agree with. But the best one is this "I thus believe that in the future we will reach a new “natural boundary” in mathematics, beyond (and perhaps way beyond) where we are now, but where machines are going to get stuck and where it is not viable to expend any more resources to make the next big leap. (...) I believe that the optimal thing to do (...) is to let the machines loose, see what happens, and then begin the journey to where they have stopped." https://t.co/6J55GjyaST

Denis Yaratsreposted3:58 AM

RepostedDenis Bykov@denis_bykovHey, if you've tried any of @perplexity_ai APIs (Search, Agent, Embeddings, Decision, etc.) and were disappointed by anything, please DM me! Would be glad to learn what we should do better :)

In the conversation

Mentions

PMax Snow12
Yobi AI12
PJudy Shapiro11
Topic Intelligence™️11
Perplexity9
Clem Delangue8
PClément Delangue8
OpenAI7
Google6

Where it comes from

Items

X98
TechCrunch AI9
New York Times Technology7
AlleyWatch4
TechCrunch Startups1
Brooklyn Bridge Ventures1

Based on the 120 items in this feed. Mention and source counts describe this coverage sample, not the whole ecosystem.

Explore media & content data ↗