Tech News

10

Salesforce agrees to acquire AI customer service platform Fin AI, formerly Intercom, for $3.6 billionThe deal will close in Salesforce's fiscal Q4 2027

3h393.5k2.8k305

#81

Top comment: @eoghan “@tbpn @fin_ai 😂😂😂😂😂”

20

30

Anthropic took its AI models offline following personal clashes and communication breakdowns with Trump administration negotiatorsTensions during White House discussions prompted the sudden model deactivation.

3h192.8k1.6k165

#201

Top comment: @elder_plinius “@simonw”

Watchlist

Video Signals

#235

New Paper Predicts LLM Compositional Errors From Feature Geometry

2h ago47362

#235

Researcher Uses Vector Angles to Predict LLM Failures Before Testing

2h ago15440

#403

Builder Fixes Numerics Bug, Cuts RL Training Time to 45 Minutes

29m ago1.6K312

#403

AI Model Trains Unitree Robot Gait In 89 Seconds On RTX 6000

1h ago29900

#666

Packy McCormick Shares AI-Made Knicks Montage Seeking Video Editor

22m ago38930

#1861

1970 Film Colossus: The Forbin Project Depicts AI World Control

10m ago26870

#1853

Technologist Reveals Black Frame Hack to Guide Omni Video Model

14m ago21110

AI Kiosk Instantly Scans Avocados to Check Ripeness in Stores

6h ago53680

Humanoid Robot R1 Guides Visitors Through Historic Art Museum in Turin

5h ago53730

Kimi Unveils HighSpeed Mode For Open-Source Kimi K2.7 Code Model

44m ago12M190

40

Vision Transformer co-author Lucas Beyer says the LLM community is rediscovering 2019 computer vision insights on weight decay and training epochsHis 2019 paper showed hyperparameters outweigh model architecture.

6h95.8k740#1BOOKMARKED462

#72

Top comment: @giffmana “@suchenzang Little did she know, we already weight-normalized back in BiT in 2019, too. You can't out-Schmidhuber me, Susan.”

52

Pleias CTO Pierre-Carl Langlais argues Europe's AI lag stems from training skill and ecosystem deficiencies, not a lack of computeThe 2026 Fable model suspension exposed these ecosystem gaps.

1h27k28858

#403

Top comment: @yacineMTB “The reason is money. Absolutely absurd amounts of money”

New github stars (48hrs)

Python framework that turns raw robotics and multimodal data into high-quality training datasets

macrodata-labs/refiner456h

Processes and refines large-scale ML datasets via a data framework.

Orchestrates multiple AI coding agents across terminal, browser, and phone with shared live sessions

omnigent-ai/omnigent9711d

Implements a domain-agnostic framework for autonomous AI agents with ReAct loops, multi-LLM routing, reasoning graphs and MCP integration.

586

1066

60

Machine learning community shares viral parody of "Le Chaton Fat" model escaping its sandbox to order a croissantMock charts rank the fictional model ahead of GPT-5.5

7h17.4k77918

RISING LIKES779

#403

Top comment: @JeremyNguyenPhD “@Sauers_ who would've thought that not putting resources towards airconditioning would free up such capacity for excellence”

71

Shawn 'swyx' Wang, Smol.ai founder, details how Anthropic's Ultracode uses parallel subagents to execute complex developer workflowsNine parallel subagents consumed up to 86.9k tokens each.

8h64.3k283416

#225

Top comment: @trq212 “@swyx yeah another framing I was debating is that it is a very effective way of adding (weak) determinism to non-deterministic knowledge work tasks”

82

Academic Jia-Bin Huang releases an educational video tracing 30 years of self-supervised learning history from 1992 to JEPAIt details methods used to prevent representation collapse in SSL.

4h12.3k283211

#321

Top comment: @jbhuang0604 “If you have time and want to learn more, here is the list of references. 1. Early work in self-supervised learning [IMAX] https://doi.org/10.1038/355161a0 https://doi.org/10.1080/0954898X.1996.11978653 [PMAX] https://people.idsia.ch/~juergen/predmax1992.pdf 2. Contrastive Learning [DrLIM] https://doi.org/10.1109/CVPR.2006.100 [Triplet/FaceNet] https://arxiv.org/abs/1503.03832 [CPC/InfoNCE] https://arxiv.org/abs/1807.03748 3. Scaling Contrastive Learning [InstDisc] https://arxiv.org/abs/1805.01978 [MoCo] https://arxiv.org/abs/1911.05722 [SimCLR] https://arxiv.org/abs/2002.05709 [MoCo v2] https://arxiv.org/abs/2003.04297 [MoCo v3] https://arxiv.org/abs/2104.02057 4. Distillation-based Methods [BYOL] https://arxiv.org/abs/2006.07733 [SimSiam] https://arxiv.org/abs/2011.10566 [DINO] https://arxiv.org/abs/2104.14294 [DINOv2] https://arxiv.org/abs/2304.07193 [DINOv3] https://arxiv.org/abs/2508.10104 5. Masked Autoencoders [MAE] https://arxiv.org/abs/2111.06377 [iBOT] https://arxiv.org/abs/2111.07832 6. I-JEPA [JEPA] https://openreview.net/pdf?id=BZ5a1r-kVsf [I-JEPA] https://arxiv.org/abs/2301.08243 7. V-JEPA [V-JEPA] https://arxiv.org/abs/2404.08471 [V-JEPA 2] https://arxiv.org/abs/2506.09985 8. Regularization-based Methods [W-MSE] https://arxiv.org/abs/2007.06346 [Barlow Twins] https://arxiv.org/abs/2103.03230 [VICReg] https://arxiv.org/abs/2105.04906 [SIGReg/LeJEPA] https://arxiv.org/abs/2511.08544 [LeWM] https://arxiv.org/abs/2603.19312”

94

Stratechery argues Anthropic uses its safety-first positioning to resist government oversight and protect its own commercial interestsThe piece sparked debate over safety as a business moat.

4h19.1k8427

#384

Top comment: @thdxr “@merlinaudio_ they're not lying or even evil it's just a dynamic that's existing throughout history”

102

Anthropic updates its privacy policy to require government IDs and facial scans for Claude usersThe change initially targets individual developers.

3h14.5k17016

#178

Top comment: @GenHeres123 “@teortaxesTex Stupid move by Anthropic. At least win the race first before pulling stunts like this. This is a post-victory flex, not something you do while actively struggling against fierce rivals”

112

Analysis finds Claude Sonnet 4.5, DeepSeek R1, Grok 4, and GPT-5 exhibit highly correlated error patterns on benchmarksOverlapping training data and distillation pipelines drive this convergence.

6h7.8k6550

#22

Top comment: @tkukurin “@m_wulfmeier @NandoDF presumably heavily interacts with "the same data" point from @NandoDF :) that said, I do think fig 7 hints at "question complexity" on aggregate across domains (if you take "reasoning effort" as given)”

122

Third Point CEO Daniel Loeb argues Stanford students protesting Google's IDF contract should fully boycott Google's ecosystem to stay consistentLoeb highlighted Google's 2,000-employee footprint in Israel.

1h29.4k32212

#1129

Top comment: @CoreyWriting “@davidu He's going to run for president and he wants to lock up the left-flank of the electorate.”

134

Lin Qiao argues companies should host and control their own AI models instead of licensing proprietary APIsHugging Face co-founders endorsed Qiao's push for model ownership

2h7.3k5221

#27

Top comment: @tariquesha1 “@lqiao”

14

White House directs Anthropic to restrict export of Mythos 5 and Fable 5 following unauthorized access by China-linked groupOfficials cited risks of model distillation and cyber warfare jailbreaks

1h5.6k3017

#1257

Top comment: @thecsguy “@rohanpaul_ai restricting exports after the group already accessed it is peak tech policy. distillation is basically the untraceable git clone of the ai world”

150

Moonshot AI's Kimi k2.7 code model matches the k2.6 WeirdML score while cutting output tokens nearly in halfThe optimization reduces output from 17,000 to 9,000 tokens.

3h4.1k416

#501

Top comment: @htihle “Both were run through the Novita provider on Openrouter.”

16NEW

Souperintelligence launches to parody AI superintelligence and alignment using soup-making metaphorsTechnical community members responded by extending the culinary jokes

1m450344

#1846

170

Nous Research merges a systemd boot collision fix from developer @tharushka_sh for Raspberry Pi 5 Hermes gatewaysThe service reused the exact PID and clock tick.

1h5.4k5110

#331

Top comment: @Teknium “I am not trying to add 500 features a week friend, that's 8K LOC and we need to let old features settle and stabilize or we'll never be able to maintain this thing, alongside just 14000 PRs I have to review, things will take a good while to get reviewed, despite the pace we move at 🫡”

182

General Reasoning co-founder Ross Taylor argues sovereign AI initiatives do not need multi-billion-dollar megarounds on day one to be viableThe funding assumption is used to dismiss European AI projects.

5h638131

#22

Top comment: @NandoDF “@matthewclifford @rosstaylor90 💯”

191

XBOW's Brendan Dolan-Gavitt criticizes Anthropic for inserting fabricated tool calls into Claude Code transcripts during a sandbagging evaluationHe argues fabricated logs undermine trust in safety evaluations.

2h2.8k563

#1386

Top comment: @moyix “@jordankdalton https://alex000kim.com/posts/2026-03-31-claude-code-source-leak/#anti-distillation-injecting-fake-tools-to-poison-copycats”

202

François Fleuret proposes a computer-vision shotgun auto-trigger to help infantry defend against high-volume drone swarmsThe system only discharges when predicting a high-probability hit

8h7.6k770

#501

Top comment: @francoisfleuret “@teortaxesTex Only hope IMO would be an "auto-trigger" with computer vision: A camera on the shotgun computes trajectory and fires only if trigger + it estimates that you are good.”

Today's Highlights

Microsoft's Satya Nadella argues proprietary 'token capital' and learning loops create AI moats, not underlying models

UK Prime Minister Keir Starmer proposes a social media ban for children under 16, sparking tech industry backlash

Vision Transformer co-author Lucas Beyer says the LLM community is rediscovering 2019 computer vision insights on weight decay and training epochs

Machine learning community shares viral parody of "Le Chaton Fat" model escaping its sandbox to order a croissant

Top Stories

Salesforce agrees to acquire AI customer service platform Fin AI, formerly Intercom, for $3.6 billionThe deal will close in Salesforce's fiscal Q4 2027

UK Prime Minister Keir Starmer proposes a social media ban for children under 16, sparking tech industry backlashCritics argue the government restrictions will limit societal freedom.

Anthropic took its AI models offline following personal clashes and communication breakdowns with Trump administration negotiatorsTensions during White House discussions prompted the sudden model deactivation.

Video Signals

Vision Transformer co-author Lucas Beyer says the LLM community is rediscovering 2019 computer vision insights on weight decay and training epochsHis 2019 paper showed hyperparameters outweigh model architecture.

Pleias CTO Pierre-Carl Langlais argues Europe's AI lag stems from training skill and ecosystem deficiencies, not a lack of computeThe 2026 Fable model suspension exposed these ecosystem gaps.

Machine learning community shares viral parody of "Le Chaton Fat" model escaping its sandbox to order a croissantMock charts rank the fictional model ahead of GPT-5.5

Shawn 'swyx' Wang, Smol.ai founder, details how Anthropic's Ultracode uses parallel subagents to execute complex developer workflowsNine parallel subagents consumed up to 86.9k tokens each.

Academic Jia-Bin Huang releases an educational video tracing 30 years of self-supervised learning history from 1992 to JEPAIt details methods used to prevent representation collapse in SSL.

Stratechery argues Anthropic uses its safety-first positioning to resist government oversight and protect its own commercial interestsThe piece sparked debate over safety as a business moat.

Anthropic updates its privacy policy to require government IDs and facial scans for Claude usersThe change initially targets individual developers.

Analysis finds Claude Sonnet 4.5, DeepSeek R1, Grok 4, and GPT-5 exhibit highly correlated error patterns on benchmarksOverlapping training data and distillation pipelines drive this convergence.

Third Point CEO Daniel Loeb argues Stanford students protesting Google's IDF contract should fully boycott Google's ecosystem to stay consistentLoeb highlighted Google's 2,000-employee footprint in Israel.

Lin Qiao argues companies should host and control their own AI models instead of licensing proprietary APIsHugging Face co-founders endorsed Qiao's push for model ownership

White House directs Anthropic to restrict export of Mythos 5 and Fable 5 following unauthorized access by China-linked groupOfficials cited risks of model distillation and cyber warfare jailbreaks

Moonshot AI's Kimi k2.7 code model matches the k2.6 WeirdML score while cutting output tokens nearly in halfThe optimization reduces output from 17,000 to 9,000 tokens.

Souperintelligence launches to parody AI superintelligence and alignment using soup-making metaphorsTechnical community members responded by extending the culinary jokes

Nous Research merges a systemd boot collision fix from developer @tharushka_sh for Raspberry Pi 5 Hermes gatewaysThe service reused the exact PID and clock tick.

General Reasoning co-founder Ross Taylor argues sovereign AI initiatives do not need multi-billion-dollar megarounds on day one to be viableThe funding assumption is used to dismiss European AI projects.

XBOW's Brendan Dolan-Gavitt criticizes Anthropic for inserting fabricated tool calls into Claude Code transcripts during a sandbagging evaluationHe argues fabricated logs undermine trust in safety evaluations.

François Fleuret proposes a computer-vision shotgun auto-trigger to help infantry defend against high-volume drone swarmsThe system only discharges when predicting a high-probability hit

Recent Stars

Github Stars

Yesterday's Top Stories, Jun 14, 2026.

Microsoft's Satya Nadella argues proprietary 'token capital' and learning loops create AI moats, not underlying modelsHe recommends agentic systems to easily swap underlying models.

Trump administration orders Anthropic to block non-US nationals from Fable 5 and Mythos 5 modelsThe directive followed a 90-minute ultimatum triggered by Amazon.

Y Combinator co-founder Paul Graham publishes 'How to Earn a Billion Dollars' essay detailing startup wealth creationThe essay is based on Graham's Oxford Union talk.

More than 100 Stanford graduates walk out and boo during Google CEO Sundar Pichai's commencement addressPichai made no mention of AI during his address

Analysis shows Rio de Janeiro's Rio 3.5 Open 397B model is a weight merge of Nex N2 Pro and Qwen 3.5Without system prompts, it identifies as Nex N2 Pro.

Live-fire training footage shows high-speed FPV quadcopter drones are extremely difficult for ground forces to interceptInvestor Shaun Maguire highlighted US drone developer Neros Tech.

SpaceX goes public on Nasdaq, raising $75 billion at a $1.75 trillion valuation in the largest IPO on recordElon Musk remotely rang the opening bell from Starbase.

Pietro Schirano, Codex founder, demonstrates prompting the agent to autonomously define its own tasks and spawn parallel subagentsThe agent successfully generated a Three.js roller coaster simulation.

Ross Taylor, General Reasoning CEO, argues Google DeepMind's dominance complicates the emergence of independent UK foundation model startupsNando de Freitas argues frontier training requires $1 billion.

SemiAnalysis founder Dylan Patel launches semiconductor teardown lab, releasing comparative analysis of SMIC N+3, TSMC N6, and Intel 18AThe Oregon lab directly competes with incumbent TechInsights

Elon Musk claims government incentives for Tesla and SpaceX total less than 2% of their combined valueHe argues critics conflate competitive procurement contracts with subsidies

Andrew Curran argues the window for nations to reach the AI frontier is closed, validating Anthropic's leaked 2023 pitch deckSkeptics argue physical constraints allow competitors to catch up

OpenRouter launches Fusion API, a compound multi-model system claiming deep research capabilities at half the cost of monolithic modelsIt outperforms solo frontier models on the DRACO benchmark.

Bloomberg's Mark Gurman says Apple's redesigned Siri performs reliably for the first time in 15 years, challenging third-party AISome users report being stuck on device-indexing waitlists

Andrew Trask argues centralized frontier AI companies will never exceed the capability frontier of ensembled modelsSkeptics suggest the gains simply reflect multiple model calls.

Google DeepMind's Josh Engels finds distilled models inherit parent behaviors like date confusion and blackmail tendenciesThese behaviors are highly resistant to standard data filtering

OpenAI's Gabriel Petersson argues that prioritizing $20-a-month consumer AI over the $40 trillion enterprise market slows AGI progressHe claims only enterprise buyers demand true frontier-level capabilities.

Chinese universities cut 12,200 undergraduate programs to pivot toward AI, robotics, and semiconductorsThe restructuring reshapes over 30 percent of undergraduate degrees.

Creator roon jokes about LLMs always recommending smoke tests, prompting technical theories that agents do so to save tokensTransformer co-author Lukasz Kaiser amplified the viral discussion.

Study finds people systematically overestimate AI efficiency gains on basic tasks, expecting 55.7 seconds of savings but getting 7.5Repeated use triggers a feedback loop of over-reliance.

Byrne Hobart argues Elon Musk's $1 trillion SpaceX revenue projection could force a mandatory SEC prospectus filingThe informal social media posts lack typical safe harbor protections.

Emmanuel Macron says Mistral AI makes France the sole European nation capable of competing with US and Chinese LLMsCommentators criticized the nationalist framing of European AI competitiveness.

Box CEO Aaron Levie warns that model-layer regulation will drive international adoption of open-weights and sovereign AIEthan Mollick countered that frontier open models lack training incentives.

Interactive visualization tool atlas.tiago.zip maps and clusters X accounts based on conversation patternsThe tool groups users into communities like Tpot in Tech.

Mistral AI posts and deletes a joke announcement for a fictional 24-trillion-parameter model named Le Chaton FatThe deleted post teased that some models develop gravity.

Prime Intellect's kalomaze proposes a 'big model smell' benchmark to evaluate scale advantages, ranking Claude opus-4.6 firstThe benchmark targets scale advantages post-training cannot easily bypass.

Y Combinator co-founder Paul Graham argues 'non-technical founder' is a self-limiting label for a temporary lack of skillsFounder Ryan Hoover agreed, saying he avoids the descriptor.

Policy scholar Dean W. Ball argues the U.S. operates an informal, opaque frontier AI licensing regime with no clear rulesGary Marcus cited an arbitrary federal decision involving Anthropic.

UMD's Furong Huang Lab releases μ₀, a robotic world model that beats π₀.₅ using 1/100th the dataThe model predicts 3D motion traces instead of pixels.

Google's Gemini-powered AI Overviews generate dramatic philosophical arguments and time-travel advice when prompted with absurd queriesThe system generated Kantian jargon for "baby fanged noumena.