Tech News

Today's Highlights

In case you missed it

Midjourney launches medical hardware division, unveiling an ultrasound scanner to replace MRIs with 60-second full-body scans

JA#173|@JACHIAM0

DA#117|@DAVIDSHOLZ

FC#77|@FCHOLLET

LB#72|@GIFFMANA

EM#52|@ELONMUSK

#1 VIEWED

Bun creator Jarred Sumner says upcoming Bun release runs React Compiler natively in Rust, beating Babel by 19.8x

JS#1502|@JARREDSUMNER

KA#403|@YACINEMTB

#1 BOOKMARKED

154

THUDM open-sources Slime, the reinforcement learning scaling framework used to train Zhipu AI's GLM-5.2 model

EC#919|@ETHANCABALLERO

T(#501|@TEORTAXESTEX

RA#218|@AGARWL_

RISING LIKES

106

Prime Intellect's Will Brown says third-party vendors bypass distillation bans by selling competitor-generated training data to major labs

XL#1855|@XLR8HARDER

WB#573|@WILLCCBB

NL#80|@NATOLAMBERT

Top Stories

Posts:—|Clusters:—|Next Crawl:—

11

EU issues tenders for seven data centers with up to 40,000 GPUs, scaling back sovereign infrastructure plansObservers say the downsized footprint favors smaller AI models.

2h|11.1k10211

XL#1855|@XLR8HARDER

FB#1778|@XEOPHON

SK#667|@SEBKRIER

T(#501|@TEORTAXESTEX

Top post: @dioscuri “@sebkrier I believe the approved term is bijou”

21

THUDM open-sources Slime, the reinforcement learning scaling framework used to train Zhipu AI's GLM-5.2 modelIt merged over ten expert models in two days

7h|22.8k315#1BOOKMARKED154

EC#919|@ETHANCABALLERO

T(#501|@TEORTAXESTEX

RA#218|@AGARWL_

Top post: @teortaxesTex “@lu_sichu sorry for deleting had to check it https://huggingface.co/blog/zai-org/glm-52-blog”

Digg Deeper question: “What's the TLDR of what this does in plain english?”

32

Bun creator Jarred Sumner says upcoming Bun release runs React Compiler natively in Rust, beating Babel by 19.8xUsers invoke the compiler via the `--react-compiler` build flag.

6h|#1VIEWS72.4k#1LIKED69455#1COMMENTS38

JS#1502|@JARREDSUMNER

KA#403|@YACINEMTB

Top post: @jarredsumner “It runs the same test suite as the upstream react compiler version with tests for flow type and meta’s intl library skipped”

Watchlist

Video Signals

CH#1360|@KIMMONISMUS

WoW Private Server Runs 1,800 AI Bots With DeepSeek Chat

4h ago|Views 490KLikes 6.2KBookmarks 1.5K

Midjourney Unveils Scanner, Anduril Wins USAF Contract Amid Tech Wins

2h ago|Views 15KLikes 174Bookmarks 28

EL#684|@OMARSAR0

New YouTube Notetaker Skill Generates Slides, Notes, and Transcripts From Videos

2h ago|Views 7.1KLikes 46Bookmarks 75

FD#809|@FDELLAERT

Crysta AI 3D Prints Gaussian Splat Bee Model in Resin Cube

6h ago|Views 69KLikes 0Bookmarks 0

T(#501|@TEORTAXESTEX

Sberbank Showcases Green Humanoid Robot Powered by Qwen-4B AI

2h ago|Views 5.3KLikes 35Bookmarks 10

BR#1861|@BRIANROEMMELE

SqueezLabs Unveils Hand-Cranked CrankGPT For Off-Grid AI Inference

2h ago|Views 12KLikes 77Bookmarks 9

AK#1029|@ALEXRKONRAD

Mutiny CEO Pivots Startup to AI Sales Agent After Firing Customers

1h ago|Views 637Likes 8Bookmarks 0

BS#1756|@BILAWALSIDHU

Arcturus Advances 4D Gaussian Splatting For Immersive Sports Viewing

1h ago|Views 270KLikes 21Bookmarks 10

OpenAI Codex Autonomously Sets Up Google Cloud APIs in Browser

1h ago|Views 30KLikes 0Bookmarks 0

HS#800|@HARRYSTEBBINGS

US Bans Anthropic AI Model In First Capability-Based Regulation

28m ago|Views 1.6KLikes 3Bookmarks 1

40

Google DeepMind's Andreas Kirsch argues AI researchers can no longer ignore the downstream consequences of advancing model capabilitiesKirsch cites a lack of independent oversight and opaque deals.

5h|2.8k212

GK#1970|@GENEKOGAN

AK#271|@BLACKHC

SC#216|@SCYCHAN_BRAINS

MB#57|@MILES_BRUNDAGE

Top post: @genekogan “@Miles_Brundage It doesn’t seem like many people work on capability and safety at the same time. I kind of think most people understand they are not the difference maker on whether the consequences turn out good or bad. If they genuinely think it’s the latter, they stop working on capability.”

53

Jürgen Schmidhuber argues his Munich team's 1991 papers introduced the foundational concepts for modern transformers, network distillation, and pre-trainingSakana AI's hardmaru says the work influenced his research.

7h|37.8k27172

JS#86|@SCHMIDHUBERAI

HA#17|@HARDMARU

Top post: @hardmaru “In 1991, the foundations for Transformers, Pre-training, Distillation, and World Models were already being built. These helped shape my own thinking, from my time at Google Brain to our Recursive Self-Improvement (RSI) work at @SakanaAILabs today. 🧠🗼 https://people.idsia.ch/~juergen/ai-boom-roots-munich-1991.html 👇”

60

AI2's Nathan Lambert argues that restricting open-source AI threatens American values, transparency, and innovationFlorian Brand warned that restrictions enable monopolistic rent-seeking.

1h|11.8k7712

FB#1778|@XEOPHON

AC#682|@ANDREWCURRAN_

NL#80|@NATOLAMBERT

Top post: @natolambert “@kevinsxu Also on @interconnectsai and interconnected (yes, overlap)”

76

Ukraine launches TrophyLab, a secure platform for sharing vulnerability analyses and data on captured Russian military hardwareAuthorized partners can also request physical weapon samples

2h|7.1k549

JW#710|@WOLFEJOSH

Top comment: @lizardmech “@FedorovMykhailo You should post a tank one publically as a teaser, the internet loves secret tank details and t-90m armor is harmless information for everyone besides the company trying to sell t-90m.”

81

Prime Intellect's Will Brown says third-party vendors bypass distillation bans by selling competitor-generated training data to major labsThe workaround allows indirect training on competitor model outputs.

4h|7.3k10612

RISING LIKES106

XL#1855|@XLR8HARDER

WB#573|@WILLCCBB

NL#80|@NATOLAMBERT

Top post: @natolambert “It'll come down to, if the U.S. labs don't want distillation they shouldn't have an API. Seems like eventually they'll do this for some models, and that's their choice to make. More onerous regulation wont really work and will hurt startups in the US.”

95

Noam Shazeer's departure from Google raises doubts about Gemini Pro's future as OpenAI recruits pre-training talentOpenAI is hiring London-based pre-training researchers for model scaling

5h|31.1k26163

FASTEST CLIMBING+5

BR#1879|@BINDUREDDY

SL#1498|@SAMUELMLSMITH

Top post: @bindureddy “Noam Shazeer leaving Google doesn’t bode well Hopefully Gemini will continue to be a credible model family Flash is a good model but Pro is still up in the air 🤷‍♀️”

101

Harry Stebbings, 20VC founder, predicts consumer tech companies valued over $10B will spend $25M annually on sports partnershipsAlexis Ohanian agreed with the predicted marketing shift.

5h|19.1k11628

HS#800|@HARRYSTEBBINGS

AO#636|@ALEXISOHANIAN

Top post: @alexisohanian “@HarryStebbings 😏😏😏”

111

Teortaxes argues Zhipu AI and Tsinghua University will build a competitive Chinese AI ecosystem within a yearOpen-source contributor Stella Biderman agreed with the assessment.

7h|21.2k17218

T(#501|@TEORTAXESTEX

SB#289|@BLANCHEMINERVA

Top post: @teortaxesTex “Come to think of it, Zhipu vs DeepSeek is pretty much Tsinghua academics vs Peking University competitive programmers. Whom would you ask to do the best research? Whom to design the most performant algorithms? Of course they're all great researchers and engineers, but…”

124

High-achieving South Korean engineering students increasingly choose elite Chinese universities over domestic optionsThe shift reverses historical patterns of China recruiting Korean talent

6h|7.4k7814

DA#1469|@DAVIDLEE

T(#501|@TEORTAXESTEX

Top post: @davidlee “If true, this would be a great trend for Korea. Exposing the best young talent to a different mindset towards engineering, building, experimentation washes away the value that failure is shame”

133

GLM 5.2 (max) scores 70.1% on WeirdML benchmark, requiring 22,000 output tokens per runA lower configuration yields 67.3% accuracy using 12,000 tokens.

5h|8.2k517

FB#1778|@XEOPHON

T(#501|@TEORTAXESTEX

Top post: @teortaxesTex “off by 0.1%. I'm getting rusty Yes, 7 months of a gap. Of course GLM 5.2 is a better model than Gemini 3 Pro in most real use cases. WeirdML is a benchmark which profoundly separates Western and Eastern training culture… or used to.”

141

Elie Bakouch and Meta's Lucas Beyer argue a hypothetical AI scaling challenge would reward pre-existing infrastructure over pure capabilityTeams must use existing codebases instead of starting fresh

7h|2.4k100

EL#1136|@ELIEBAKOUCH

MH#994|@MATTHEN2

LB#72|@GIFFMANA

Top post: @matthen2 “@eliebakouch google people would just complain about not having TPUs 😅”

154

Resurfaced demo shows Midjourney founder David Holz personally coded a realtime browser-based 3D visualizationThe system runs 3D visualizations without external plug-ins.

8h|3.8k394

OC#1277|@OLIVERCAMERON

RS#447|@SCOBLEIZER

Top post: @Scobleizer “@olivercameron Totally!”

167

Midjourney founder David Holz jokes about generative AI cleansing social media timelines of negative elementsHe compared the platform's output to an idealized digital environment.

6h|4k1070

DF#1643|@DAVEFONTENOT

DA#117|@DAVIDSHOLZ

Top comment: @stevejang “@bayeslord 🫶🏽👾🙏🏽”

171

Google DeepMind's Susan Zhang highlights DeepSeek-R1's $600 billion Nvidia hit, while Will Brown points to long-term growthThe release triggered a 17 percent single-day stock decline.

8h|1.8k275

WB#573|@WILLCCBB

SZ#84|@SUCHENZANG

Top post: @willccbb “@suchenzang one rounding error among many”

182

Peter Henderson, Princeton Polaris Lab lead, proposes a temporary RL environment company to eliminate repetitive AI prose and then shut downThe venture would target writing styles commonly called AI slop

46m|957215

PH#616|@PETERHNDRSN

XM#603|@INFOXIAO

Top post: @infoxiao “@PeterHndrsn you can do it peter!!”

192

Moondream AI co-founder Vikhyat K. says GLM-5.2's two-day OPD post-training works by collapsing multiple expert variantsThis method concentrates and densifies the model's training signal

7h|1k192

VI#1630|@VIKHYATK

T(#501|@TEORTAXESTEX

Top post: @teortaxesTex “@vikhyatk @didier_lopes yeah the word "entire" was throwing me off but for OPD I guess makes sense, if they encounter no problems”

207

Lan Dao, Women's Health Fund founder, argues cheap AI scans solve diagnostic false positives through frequent monitoringAI compares consecutive scans to avoid invasive follow-up.

3h|971160

LA#1876|@AD0RNAI

XL#1855|@XLR8HARDER

Top comment: @mcahogarth “Yep, most of these things are benign or resolve themselves without intervention. That's why I say the barrier is cheap monitoring. Getting multiple MRIs to keep an eye on an anomaly is unreasonable. You have to treat everything you see as a potential problem, because no one is going to schedule and pay for a sequence of imaging. One full body scan a year? Useless. One a week? Different story. That much data might change our understanding of the body's fluctuations, also, which means those false positives might reduce, and we'll know 'this is typical of most humans, this is when it becomes actionable.' But we just don't have the data to have those insights. Yet.”

Rising Stories

Claude Distillation Win1H AGO
Shashwat Goel finds Claude Opus 4.8 topped PostTrainBench by distilling traces from Chinese models R1 and GLM
#84
743.8k
Rat Brain Breakthrough7H AGO
BioComputingCo claims its models running on lab-grown rat brain cells outperform silicon on image-processing benchmarks
#1029
2524

Recent Stars

…abs/tabpfn-extensions2H AGO
Extends TabPFN with modules for interpretability, many-class classification, embeddings, unsupervised tasks and data augmentation.
1702
308 stars
EpicGames/lore16H AGO
Implements a centralized content-addressed revision control system with fragment deduplication for large binary files.
1014
2.9k stars
METR/hawk19H AGO
Deploys Kubernetes pods to run Inspect AI evals at scale with LLM proxying, log streaming, and PostgreSQL result storage.
953
28 stars
…rinceton/ceobench-src22H AGO
Simulates 500-day startup operations to benchmark long-horizon LLM agents via business databases, tools, and market events.
802
4 stars
chrischoy/Pointcept1D AGO
Provides codebase and implementations for point cloud perception research including Point Transformers and pretraining methods.
768
3 stars

Github Stars

(7 days)

Noumena-Network/code2D AGO
Provides terminal AI coding assistant for codebase inspection, file edits, command execution and multi-step workflows.
573
1136
86 stars
…r-turner/claude-guard2D AGO
Wraps Claude Code sessions with deterministic checks on commands and writes to limit agent access.
1278
570
29 stars
omnigent-ai/omnigent5D AGO
Implements a domain-agnostic framework for autonomous AI agents with ReAct loops, multi-LLM routing, reasoning graphs and MCP integration.
586
1066
3.8k stars

Privacy Policy|Terms of Service|© 2026 Digg Inc.

Yesterday's Top Stories, Jun 18, 2026.

Posts: 4,441|Clusters: 1,439|Stories frozen at midnight PT

10

Midjourney launches medical hardware division, unveiling an ultrasound scanner to replace MRIs with 60-second full-body scansThe system uses 500,000 underwater transducers for 200-micron resolution

1d|2.3M39.1k4.3k

JC#1402|@JAMESCHAM

SI#1378|@SUZATWEET

AC#1375|@ANDERCOT

CH#1360|@KIMMONISMUS

RH#1332|@RRHOOVER

RP#1257|@ROHANPAUL_AI

SM#1225|@PITDESI

LA#1215|@SCALING01

DM#1198|@DAVEMORIN

AV#1193|@AVISCHIFFMANN

AN#1147|@ML_ANGELOPOULOS

CH#1144|@CELINEHALIOUA

TA#1140|@VICTORTAELIN

SH#1130|@SHYAMALANADKAT

MP#1028|@EASTDAKOTA

AM#1027|@ADAMMARBLESTONE

KD#904|@CSPROFKGD

PT#821|@ELDER_PLINIUS

DF#756|@DANIELLEFONG

AO#636|@ALEXISOHANIAN

AJ#617|@JM_ALEXIA

TM#613|@ISCIENCELUVR

XM#603|@INFOXIAO

GA#598|@GABRIEL1

YJ#586|@YUCHENJ_UW

B(#530|@BEFFJEZOS

ML#500|@MLEVCHIN

MS#492|@MICSOLANA

RS#447|@SCOBLEIZER

KI#446|@KIPPERRII

AM#437|@ANJNEYMIDHA

KA#403|@YACINEMTB

👩‍💻P#387|@DYNAMICWEBPAIGE

BP#381|@BORISMPOWER

JB#304|@JON_BARRON

MJ#295|@MAXJADERBERG

AG#251|@ANIMESH_GARG

LC#219|@LULUMESERVEY

JA#173|@JACHIAM0

DA#117|@DAVIDSHOLZ

FC#77|@FCHOLLET

LB#72|@GIFFMANA

EM#52|@ELONMUSK

20

OpenAI hires former White House AI adviser Dean W. Ball to lead its new Strategic Futures policy teamHe previously authored the White House AI Action Plan.

15h|954.6k8k984

V(#1939|@REACH_VB

B(#1623|@ANDERSONBCDEFG

NS#1570|@NABEELQU

JW#1566|@JIAXINWEN22

ED#1551|@ELIDOURADO

GH#1539|@GHADFIELD

TK#1513|@TOMEKKORBAK

TH#1478|@TIMHWANG

CM#1436|@CHARLIERMARSH

HB#1424|@HERBIEBRADLEY

AM#1396|@ANDREWMAYNE

CH#1360|@KIMMONISMUS

KR#1316|@KATE_ROUCH

JW#1212|@J_ASMINEWANG

MW#1179|@MCKAYWRIGLEY

BA#1115|@BAYESLORD

SH#1085|@SONYATWEETYBIRD

🎭#1014|@DEEPFATES

DR#981|@DANINTHEORY

AE#946|@ADRIENLE

MK#869|@MIKEKNOOP

JP#825|@ZHANSHENG

DF#756|@DANIELLEFONG

YS#683|@YONASHAV

AC#682|@ANDREWCURRAN_

MS#585|@MARTINSHKRELI

B(#530|@BEFFJEZOS

T(#501|@TEORTAXESTEX

JK#480|@JASONKWON

KI#446|@KIPPERRII

KA#403|@YACINEMTB

RN#363|@RICHARDMCNGO

AK#271|@BLACKHC

DK#207|@DAVIDSKRUEGER

BM#188|@BOBMCGREWAI

BB#154|@BOAZBARAKTCS

SD#104|@_SHOLTODOUGLAS

MC#101|@MARKCHEN90

JH#90|@JEREMYPHOWARD

NL#80|@NATOLAMBERT

MB#57|@MILES_BRUNDAGE

NB#43|@POLYNOAMIAL

Digg Deeper question: “Will this delay gpt 5.6?”

30

Instagram algorithm author Thomas Dimson launches In The Weights to check if your name is encoded in AI modelsThe tool measures name recognition across 14 different AI models.

13h|194.8k903207

V(#1939|@REACH_VB

LT#1719|@_LEWTUN

ND#1653|@NICKADOBOS

BR#1631|@BLAKEIR

BT#1605|@BENTOSSELL

AD#1537|@DORIALEXANDER

TD#1501|@TURTLESOUPY

SS#1471|@SYTELUS

AV#1419|@ALTRYNE

SL#1366|@LESSIN

AN#1270|@ALLENAINIE

TW#1264|@TW_KILLIAN

CY#1197|@CHHAVIYADAV_

EL#1136|@ELIEBAKOUCH

YO#1081|@YOHEINAKAJIMA

CA#1059|@CHRISALBON

DF#947|@CANONDETORTUGAS

ME#861|@MERVENOYANN

JM#792|@VENTURETWINS

DF#756|@DANIELLEFONG

AA#748|@ARYAMAN2020

PM#714|@PMINERVINI

MC#660|@MMMBCHANG

PV#647|@PETARV_93

CT#542|@CHENHAOTAN

👩‍💻P#387|@DYNAMICWEBPAIGE

AK#271|@BLACKHC

RA#218|@AGARWL_

LF#98|@LIAMFEDUS

SD#91|@SEDIELEM

LB#72|@GIFFMANA

40

Transformer co-author Noam Shazeer leaves Google to join OpenAI as lead of architecture researchHe also pioneered foundational research into mixture-of-experts systems.

1d|349.3k3.5k159

TC#1914|@TIFAFAFAFA

YP#1896|@YAMPELEG

CL#1561|@CLU_CHENG

CH#1360|@KIMMONISMUS

EN#1301|@ERICNEWCOMER

RP#1257|@ROHANPAUL_AI

KA#1213|@KALOMAZE

SA#1139|@SAMUELALBANIE

AJ#874|@ARMANDJOULIN

PD#834|@PMDDOMINGOS

PB#771|@PAULTOO

NT#673|@NICKATURLEY

SM#648|@SWAROOPRM7

RH#642|@ROMAINHUET

TM#613|@ISCIENCELUVR

GA#598|@GABRIEL1

AC#484|@ANDREW_N_CARR

RS#447|@SCOBLEIZER

BP#381|@BORISMPOWER

SY#351|@SONGLINYANG4

RI#232|@RISHIBOMMASANI

EH#193|@ERICHORVITZ

YT#160|@YITAYML

50

Zai_org launches GLM-5.2, an open-weights 744B model with a 1M context window that beats GPT-5.5 on FrontierSWEThe model is released under the MIT License.

5d|636.6k7.8k1.5k

BR#1861|@BRIANROEMMELE

FB#1778|@XEOPHON

ZE#1695|@ZEPHYR_Z9

RP#1257|@ROHANPAUL_AI

LA#1215|@SCALING01

ME#861|@MERVENOYANN

CP#855|@CHRIS_J_PAXTON

AC#682|@ANDREWCURRAN_

T(#501|@TEORTAXESTEX

RS#447|@SCOBLEIZER

KA#403|@YACINEMTB

T🪽#331|@TEKNIUM

JH#90|@JEREMYPHOWARD

60

OpenAI launches Record and Replay for Codex, allowing users to turn manual tasks into reusable, editable skillsIt automates repetitive administrative tasks like expense reports.

12h|822.7k4.7k2.1k

V(#1939|@REACH_VB

“P#1846|@PAULARAMBLES

MB#1780|@MATTHEWBERMAN

BS#1756|@BILAWALSIDHU

TS#1677|@TIMSHI_AI

AE#1416|@EMBIRICO

AM#1396|@ANDREWMAYNE

🎭#1014|@DEEPFATES

DS#975|@DANSHIPPER

MM#716|@MASCOBOT

MS#691|@MATTSHUMER_

PS#382|@STEIPETE

SY#351|@SONGLINYANG4

70

Senator Bernie Sanders proposes giving the public a 50% equity stake in AI companies exceeding $200 million in revenueTech investors warned the mandate would deter new startups

16h|901.6k#1LIKED44.1k531

BR#1861|@BRIANROEMMELE

SO#1349|@SOLEIO

JM#792|@VENTURETWINS

B(#530|@BEFFJEZOS

GT#121|@GARRYTAN

MB#57|@MILES_BRUNDAGE

EM#52|@ELONMUSK

80

OpenAI and Boston Children's Hospital use o3 Deep Research to diagnose 18 unsolved pediatric genetic casesThe findings were published in the journal NEJM AI.

15h|471.6k3.2k509

KS#1976|@THEKARANSINGHAL

CM#1436|@CHARLIERMARSH

NL#1407|@LABENZ

BM#1004|@MCKBRANDO

SR#957|@CLONEOFSIMO

TM#613|@ISCIENCELUVR

JK#480|@JASONKWON

CL#468|@CHENGLEISI

BP#381|@BORISMPOWER

NB#43|@POLYNOAMIAL

90

Claude Code releases Artifacts beta to convert active coding sessions into interactive, shareable HTML pagesThe beta is exclusive to Team and Enterprise plans.

12h|761.4k5.9k3.2k

JS#1502|@JARREDSUMNER

AC#682|@ANDREWCURRAN_

LR#543|@LENNYSAN

BC#369|@BCHERNY

AT#329|@ALEXTAMKIN

100

GLM creator Jie Tang disputes Elon Musk's Q1 2027 timeline for Chinese AI to reach Western frontier parityBenchmarks place GLM 5.2 seven months behind Claude.

1d|303.7k4k410

BR#1861|@BRIANROEMMELE

ZE#1695|@ZEPHYR_Z9

CH#1360|@KIMMONISMUS

RO#1315|@KRISHNANROHIT

LA#1215|@SCALING01

🎭#1014|@DEEPFATES

EL#684|@OMARSAR0

T(#501|@TEORTAXESTEX

KA#403|@YACINEMTB

110

US Vice President JD Vance warns that AI functions as a communist technology by enabling deep corporate and government surveillanceHe warned the technology could enable social credit systems

20h|697.7k4.6k852

MN#1716|@MENHGUIN

SB#289|@BLANCHEMINERVA

GM#178|@GARYMARCUS

120

Coinbase CEO Brian Armstrong and Elon Musk argue orbital datacenters will surpass terrestrial compute to bypass Earth's regulatory bottlenecksArmstrong proposed capping regulatory growth in Mars legal frameworks

13h|478.9k10.6k579

VP#1821|@VIJAYPANDE

DF#756|@DANIELLEFONG

@L#627|@LEVELSIO

BA#398|@BRIAN_ARMSTRONG

EM#52|@ELONMUSK

130

OpenAI's @tszzl questions the competitive risk of disclosing reinforcement learning and chain-of-thought scaling principles during the o1 launchThe debate highlights the competitive cost of publishing AI research.

6h|205.7k2.3k282

LA#1215|@SCALING01

SB#289|@BLANCHEMINERVA

EM#184|@EMOLLICK

JT#138|@MILLIONINT

RA#102|@_AROHAN_

MB#57|@MILES_BRUNDAGE

140

US Commerce Secretary Howard Lutnick warns ASML a restricted EUV lithography machine may have entered ChinaASML denies ever shipping an EUV tool to China.

6h|264.2k2k419

MN#1716|@MENHGUIN

ZE#1695|@ZEPHYR_Z9

AC#682|@ANDREWCURRAN_

T(#501|@TEORTAXESTEX

MB#57|@MILES_BRUNDAGE

150

Forward Future founder Matthew Berman launches Loop Library, a directory of repeatable, verified AI agent workflowsWorkflows include automated ticket-to-PR and page-load optimization loops

12h|525.2k3.7k#1BOOKMARKED6.4k

MB#1780|@MATTHEWBERMAN

MS#1765|@MATTPRD

160

xAI integrates Grok models into Databricks Agent Bricks to let enterprises build AI agents on proprietary dataThe integration maintains secure enterprise control over private data.

14h|#1VIEWS4.6M19k983

AG#1613|@ALIGHODSI

EM#52|@ELONMUSK

170

Trump administration issues 90-minute ultimatum to Anthropic, forcing global suspension of Fable 5 and Mythos 5 modelsThe order bars non-US employees from training the models.

6d|225.9k4.2k186

AK#982|@ANDREY_KURENKOV

BB#931|@WGRATHWOHL

PT#821|@ELDER_PLINIUS

SS#819|@STEVESI

180

Elon Musk argues Anthropic successfully prioritized practical, revenue-driving utility over standard AI benchmark scoresEmad Mostaque says Chinese labs similarly prioritize real-world utility

12h|171.6k2k236

ZE#1695|@ZEPHYR_Z9

CH#1360|@KIMMONISMUS

LA#1215|@SCALING01

TA#1140|@VICTORTAELIN

T(#501|@TEORTAXESTEX

EM#171|@EMOSTAQUE

190

White House and Anthropic collaborate on a standardized technical framework to benchmark AI jailbreak severityThe framework could restore access to Mythos and Fable.

10h|112k1.2k142

AS#1809|@ALEXSTAMOS

TH#1478|@TIMHWANG

AC#682|@ANDREWCURRAN_

200

Paul Graham says the next major computing form factor is unknown but will seem obvious in retrospectMinh Nhat Nguyen mocked the prediction with a meme.

15h|218.7k3.4k426

MN#1716|@MENHGUIN

DF#1643|@DAVEFONTENOT

210

A resurfaced post shows Midjourney co-founder David Holz proposed launching a biomedical imaging division to reduce scan false positivesThe March 2024 proposal targeted annual cancer scans.

15h|89.8k3.4k293

“P#1846|@PAULARAMBLES

DF#1643|@DAVEFONTENOT

LB#773|@REMEMBERLENNY

DE#274|@DEEDYDAS

220

Anthropic's Chris Ciauri says blocked frontier models Mythos and Fable 5 will regain availability in South Korea within daysAccess restoration follows positive G7 negotiations with President Trump

18h|191.6k2.2k210

ZE#1695|@ZEPHYR_Z9

CH#1360|@KIMMONISMUS

AC#682|@ANDREWCURRAN_

230

Accenture shares plunge up to 20% after lowered revenue forecast blames AI disruption of consulting workflowsThe selloff dragged down rivals Infosys, Tata, and Cognizant

16h|221.8k1.7k221

RP#1257|@ROHANPAUL_AI

WM#520|@WILLMANIDIS

GM#178|@GARYMARCUS

240

Open-source builder _xjdr launches Noumena Code to prevent Git merge conflicts in multi-agent workflowsHermes creator Teknium says standard Git handles agent workflows easily

17h|107.7k1.1k747

XL#1855|@XLR8HARDER

🎭#1014|@DEEPFATES

250

a16z's independent media division releases its first annual report detailing its direct-to-audience content strategy for foundersThe division operates separately from the core venture firm.

16h|378.3k584457

SW#1354|@SARAHDINGWANG

SS#819|@STEVESI

LR#543|@LENNYSAN

ET#524|@ERIKTORENBERG

260

Richard Ngo argues the AI safety community's narrow focus on AGI and recursive self-improvement constrains thinking at labsLeo Gao countered that AGI concern is still essential.

15h|111k895412

C❤️‍🔥#1909|@CLAREJTBIRCH

B(#530|@BEFFJEZOS

LG#469|@NABLA_THETA

RN#363|@RICHARDMCNGO

SZ#84|@SUCHENZANG

270

Valar Atomics takes its Ward 250 advanced nuclear reactor critical for the first timeThe startup plans to deliver power by July 4.

7h|115.1k1.7k64

CP#1748|@2112POWER

KC#777|@KIMMAICUTLER

SM#550|@SHAUNMMAGUIRE

B(#530|@BEFFJEZOS

BA#398|@BRIAN_ARMSTRONG

280

Founders Fund releases final "MAFIA" episode featuring Sam Altman, Elon Musk, and Peter Thiel playing the social deduction gameThe 24-minute production features group discussions and card reveals

13h|144.5k852147

BS#1756|@BILAWALSIDHU

TS#1189|@TRAESTEPHENS

HE#708|@HEINRICHKUTTLER

DE#675|@ZEBULGAR

MS#492|@MICSOLANA

RP#395|@TYPESFAST

290

Poolside AI releases Laguna M.1, a 225B-A23B open-weights MoE model with a 256K context windowThe model is licensed Apache 2.0 for codebase exploration.

15h|58.7k640125

FB#1778|@XEOPHON

LT#1719|@_LEWTUN

EL#1136|@ELIEBAKOUCH

DF#756|@DANIELLEFONG

J⧉#674|@REPLIGATE

WB#573|@WILLCCBB

KA#403|@YACINEMTB

C🤗#109|@CLEMENTDELANGUE

300

OpenAI releases beneficial reinforcement learning research evaluating GPT-5.5 and Claude Opus 4.7 on persistent alignmentThe framework tests how models maintain safe behavior under pressure.

9h|68.9k473108

KS#1976|@THEKARANSINGHAL

RP#1257|@ROHANPAUL_AI

LA#1215|@SCALING01

LK#722|@LUKASZKAISER

EM#184|@EMOLLICK