Tech News

Today's Highlights

In case you missed it

Nobel laureate and AlphaFold lead John Jumper departs Google DeepMind to join rival AI lab Anthropic

EP#151|@ETHANJPEREZ

SZ#84|@SUCHENZANG

DR#50|@ROYDANROY

ZL#31|@ZACHARYLIPTON

DH#29|@DEMISHASSABIS

It’s a new day at Digg. Fresh stories are clustering now. 25% complete.

Top Stories

Posts:—|Clusters:—|Next Crawl:—

11

Parody website InTheWeights.com launches to rank tech figures on their hypothetical presence in fictional future AI model weightsAlexis Ohanian and Shawn Wang shared their mock profiles.

4h|13.9k7411

PK#1310|@PATRICKKIDGER

AO#636|@ALEXISOHANIAN

JH#90|@JEREMYPHOWARD

Top post: @jeremyphoward “Battle of the Jeremy Howards!”

21

THUDM open-sources Slime, an LLM post-training framework for reinforcement learning scaling and online preference optimizationIt completed GLM-5.2 post-training in two days.

4h|23795

SI#1378|@SUZATWEET

T(#501|@TEORTAXESTEX

JH#90|@JEREMYPHOWARD

Top comment: @teortaxesTex “@didier_lopes > The entire OPD post-training of GLM-5.2 took on this slime platform took ~2 days. what”

31

Miles Brundage and Susan Zhang argue against nationality-based restrictions on AI talent, urging a focus on basic securityThe debate focused on national security and corporate espionage.

5h|2.7k362

SZ#84|@SUCHENZANG

MB#57|@MILES_BRUNDAGE

Top post: @Miles_Brundage “@suchenzang @willdepue (See my second tweet in the thread for many things we can do besides “discriminate against ppl” which I agree would be counterproductive even by the lights of the US competition angle (which is not the only relevant one! I have written on US-China safety collab for example)”

Watchlist

Video Signals

CH#1360|@KIMMONISMUS

Trump Says Anthropic No Longer Poses National Security Threat

19m ago|Views 12KLikes 33Bookmarks 4

AI Engineer Event Uses Polymarket To Value World Cup Suite For USA Match

5h ago|Views 3.3KLikes 4Bookmarks 0

RV|@RONALD_VANLOON

Toyota Unveils CUE7 AI Robot That Shoots Basketball Like a Pro

3h ago|Views 339Likes 2Bookmarks 0

RV|@RONALD_VANLOON

AI App Remembers Networking Context To Prevent Lost Revenue

3h ago|Views 263Likes 9Bookmarks 0

RV|@RONALD_VANLOON

Cognify AI Implants Tailored Memories to Replace Prison Sentences

3h ago|Views 199Likes 2Bookmarks 1

RV|@RONALD_VANLOON

NVIDIA Demonstrates Physical AI Skills in Large-Scale Robot Simulations

5h ago|Views 272Likes 2Bookmarks 0

RV|@RONALD_VANLOON

KNOWIN Unveils AI-Powered Humanoid Robot for Household Tasks

2h ago|Views 229Likes 0Bookmarks 0

RV|@RONALD_VANLOON

PADBOT S5 Robot Patrols Outdoors With 25-Degree Climb And 3D Mapping

8m ago|Views 27Likes 1Bookmarks 0

41

Representative Ro Khanna defends proposed 2% California billionaire tax against VC warnings of capital flightKhanna cited California's 85% share of Q1 venture funding.

3h|1.1k160

Top post: @m2jr “You can’t spell pROpaganda without Ro.”

5

Meta's Lucas Beyer and François Fleuret joke that AI writing CUDA kernels is like an eager dogEngineer kache called AI-assisted GPU programming a daily routine

3h|3k50

KA#403|@YACINEMTB

LB#72|@GIFFMANA

Top post: @yacineMTB “@francoisfleuret every day”

Recent Stars

THUDM/slime3H AGO
Integrates Megatron training with SGLang rollouts for LLM post-training and RL scaling.
90
6.5k stars
datalab-to/lift13H AGO
Extracts structured JSON from PDFs/images using a 9B vision model and schema-constrained decoding.
1400
98 stars
omnigent-ai/omnigent14H AGO
Implements a domain-agnostic framework for autonomous AI agents with ReAct loops, multi-LLM routing, reasoning graphs and MCP integration.
684
586
1066
4.1k stars
…otianxt/mimestreamctl20H AGO
CLI exposing Mimestream email ops (read/reply/move/unsubscribe) to agents/scripts via AppleScript and accessibility APIs on macOS.
526
3 stars
…abs/tabpfn-extensions23H AGO
Extends TabPFN with modules for interpretability, many-class classification, embeddings, unsupervised tasks and data augmentation.
1702
308 stars

Github Stars

(7 days)

omnigent-ai/omnigent14H AGO
Implements a domain-agnostic framework for autonomous AI agents with ReAct loops, multi-LLM routing, reasoning graphs and MCP integration.
684
586
1066
4.1k stars
Noumena-Network/code3D AGO
Provides terminal AI coding assistant for codebase inspection, file edits, command execution and multi-step workflows.
573
1136
86 stars
…r-turner/claude-guard3D AGO
Wraps Claude Code sessions with deterministic checks on commands and writes to limit agent access.
1278
570
29 stars

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

Yesterday's Top Stories, Jun 19, 2026.

Posts: 3,433|Clusters: 1,214|Stories frozen at midnight PT

10

Nobel laureate and AlphaFold lead John Jumper departs Google DeepMind to join rival AI lab AnthropicDeepMind CEO Demis Hassabis praised Jumper's historic contributions.

14h|2.2M19.1k2.4k

LA#1783|@LIORONAI

FB#1778|@XEOPHON

BH#1762|@BENHYLAK

ZE#1695|@ZEPHYR_Z9

JA#1672|@JULIUSADML

RM#1575|@RACHELMETZ

DT#1422|@DIVY93T

CH#1360|@KIMMONISMUS

RP#1257|@ROHANPAUL_AI

LA#1215|@SCALING01

RS#741|@ZIV_RAVID

JW#710|@WOLFEJOSH

AC#682|@ANDREWCURRAN_

YJ#586|@YUCHENJ_UW

XW#559|@XIAOLONW

KC#552|@KYLECRANMER

CT#542|@CHENHAOTAN

T(#501|@TEORTAXESTEX

CL#468|@CHENGLEISI

QG#364|@QUANQUANGU

NB#353|@NATHANBENAICH

AK#271|@BLACKHC

SF#180|@STANISLAVFORT

EP#151|@ETHANJPEREZ

SZ#84|@SUCHENZANG

DR#50|@ROYDANROY

ZL#31|@ZACHARYLIPTON

DH#29|@DEMISHASSABIS

Digg Deeper question: “On a 1-10, how harmful are these recent departures to Google's AI ambitions? Consider their existing known staff of researchers and key AI talent.”

20

Elon Musk argues interstellar travel will shift economic metrics from currency to mass and energyCasey Handmer outlined a roadmap to scale antimatter production.

12h|#1VIEWS30.5M#1LIKED301.3k#1BOOKMARKED15.8k

LA#1215|@SCALING01

CH#880|@CJHANDMER

DF#756|@DANIELLEFONG

MS#585|@MARTINSHKRELI

B(#530|@BEFFJEZOS

EM#52|@ELONMUSK

Digg Deeper question: “Approximately what percentage of Elon's predictions actually end up being true?”

30

Former OpenAI engineer Will Depue claims China has full access to OpenAI and Anthropic's internal systemsHe warns stolen architectures could power Chinese open-source models

7h|296.5k2.8k431

FB#1778|@XEOPHON

ST#1694|@_ONIONESQUE

EL#1136|@ELIEBAKOUCH

AC#682|@ANDREWCURRAN_

DA#619|@GROWING_DANIEL

FF#577|@FRANCOISFLEURET

T(#501|@TEORTAXESTEX

KA#403|@YACINEMTB

SF#180|@STANISLAVFORT

SZ#84|@SUCHENZANG

NL#80|@NATOLAMBERT

MB#57|@MILES_BRUNDAGE

40

Donald Trump says Anthropic no longer poses a national security threat after meeting with CEO Dario AmodeiThe Commerce Department restricted foreign access to Fable 5.

16h|1.9M7.5k2.6k

BR#1879|@BINDUREDDY

CH#1360|@KIMMONISMUS

RP#1257|@ROHANPAUL_AI

LA#1215|@SCALING01

AC#682|@ANDREWCURRAN_

KA#403|@YACINEMTB

Digg Deeper question: “If you had to speculate as to when the models would return, what's your best guess?”

50

US Commerce Secretary Howard Lutnick alleges an ASML EUV lithography machine reached China, which ASML deniesChinese research teams are actively developing domestic EUV prototypes

1d|140.9k1.3k139

CH#1360|@KIMMONISMUS

RP#1257|@ROHANPAUL_AI

ZN#983|@ZACHARYNADO

AC#682|@ANDREWCURRAN_

SK#667|@SEBKRIER

T(#501|@TEORTAXESTEX

KA#403|@YACINEMTB

MB#57|@MILES_BRUNDAGE

60

Thomas Dimson launches In The Weights, a retro site visualizing how strongly a name is represented in AI training dataThe search tool generates video game-style character profile cards

1d|57k28741

FO#1853|@FOFRAI

BH#1762|@BENHYLAK

PS#1578|@SKIRANO

TD#1501|@TURTLESOUPY

RO#1315|@KRISHNANROHIT

JH#1190|@MATHEMAGIC1AN

JE#974|@JOSHELMAN

FF#577|@FRANCOISFLEURET

AC#312|@_AIDAN_CLARK_

LC#219|@LULUMESERVEY

70

OpenAI's roon argues human-computer hybrids cannot compete with pure machine intelligence because biological brains are too slowCritics debated if mind uploading or consciousness validates hybrid systems.

8h|76.2k1.1k160

BA#1115|@BAYESLORD

DF#756|@DANIELLEFONG

DA#619|@GROWING_DANIEL

B(#530|@BEFFJEZOS

AM#359|@AIDAN_MCLAU

80

Jason Liu, Instructor creator on OpenAI's Codex team, launches seamless thread handoffs between local and remote hostsThe update automatically transitions active sessions without manual setup.

10h|183.6k1.9k521

V(#1939|@REACH_VB

MC#660|@MMMBCHANG

TI#611|@THSOTTIAUX

MC#101|@MARKCHEN90

90

OpenHands co-founder Graham Neubig says GLM-5.2 is the first open model capable of replacing closed-source models in his coding workflowsThe 744B model features a one-million-token context window

6d|52.4k614194

HC#698|@HWCHASE17

T(#501|@TEORTAXESTEX

C🤗#109|@CLEMENTDELANGUE

CS#107|@CHRSZEGEDY

JH#90|@JEREMYPHOWARD

TW#27|@THOM_WOLF

100

Leaked Google DeepMind report reveals internal discontent as its AI models slip to fifth place on industry benchmarksEmployees worry leadership is conceding the race to AGI

14h|179.9k1.3k199

ZE#1695|@ZEPHYR_Z9

FC#865|@ALTIMOR

DF#756|@DANIELLEFONG

AC#682|@ANDREWCURRAN_

B(#530|@BEFFJEZOS

110

Engineer Boris Cherny uses Claude Code to analyze and attempt to decipher the ancient script Linear ACherny requested academic peer review to validate the translation.

10h|250k1.5k740

DS#975|@DANSHIPPER

BC#369|@BCHERNY

RP#358|@KHOOMEIK

AC#312|@_AIDAN_CLARK_

120

Dwarkesh Podcast host Dwarkesh Patel argues AI requires up to one million times more data than humans due to sample inefficiencyResearchers agree that improving sample efficiency is the next frontier.

13h|91.8k600481

B(#1520|@BRENDANFOODY

EL#1136|@ELIEBAKOUCH

OK#200|@LATEINTERACTION

KP#129|@SIRBAYES

DP#60|@DWARKESH_SP

130

Proponents claim Midjourney's low-cost whole-body scanner overcomes diagnostic limits, but critics argue hardware upgrades cannot solve clinical failuresProponents claim the technology cuts imaging costs by 99.9%.

1d|181k1.5k214

VP#1821|@VIJAYPANDE

DA#1469|@DAVIDLEE

TA#1140|@VICTORTAELIN

DA#619|@GROWING_DANIEL

140

Midjourney launches a medical division, unveiling a conceptual 60-second full-body ultrasonic scanner to replace MRIsMidjourney plans to deploy 50,000 scanner units globally.

2d|29.2k41034

BR#1861|@BRIANROEMMELE

SR#957|@CLONEOFSIMO

JL#906|@JERRYJLIU0

BA#713|@ADCOCK_BRETT

DA#619|@GROWING_DANIEL

RS#447|@SCOBLEIZER

DA#117|@DAVIDSHOLZ

150

Analysis argues China's GLM challenges data-scarcity assumptions by matching Claude 3 Opus performance through model distillationSanctions forced labs to use consumer GPUs for RL.

7h|91.7k1.3k206

KA#1213|@KALOMAZE

CP#855|@CHRIS_J_PAXTON

B(#530|@BEFFJEZOS

T(#501|@TEORTAXESTEX

160

Figure AI founder Brett Adcock says the company's active humanoid robots now outnumber its human employeesRobot headcount will exceed 700 units by Q2 2026.

12h|172.5k3.1k218

RP#1257|@ROHANPAUL_AI

CP#855|@CHRIS_J_PAXTON

BA#713|@ADCOCK_BRETT

170

Helen Toner, Georgetown CSET interim director, says Anthropic builds advanced AI rapidly to act as a safety vanguardShe compared rival AI systems to dangerous forest entities.

11h|119.3k1.1k339

SA#1714|@SJGADLER

AC#484|@ANDREW_N_CARR

NL#80|@NATOLAMBERT

180

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

23h|39.6k477240

BZ#1718|@BANGHUAZ

ST#1694|@_ONIONESQUE

EC#919|@ETHANCABALLERO

T(#501|@TEORTAXESTEX

RA#218|@AGARWL_

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

190

Jean Kaddour releases Sokoban Speedrun, an RL benchmark that fine-tunes Qwen3-4B-Instruct in 87 minutes using GRPOThe project modifies Andrej Karpathy's open-source nanochat pipeline

16h|55.7k516172

M🍥#1956|@MAYFER

JS#1687|@JSUAREZ

PM#714|@PMINERVINI

NL#80|@NATOLAMBERT

200

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.

17h|41.8k31324

XL#1855|@XLR8HARDER

FB#1778|@XEOPHON

LA#1215|@SCALING01

SK#667|@SEBKRIER

T(#501|@TEORTAXESTEX

210

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

18h|53.3k31175

MC#1529|@MATTHEWCLIFFORD

DF#756|@DANIELLEFONG

JW#710|@WOLFEJOSH

220

Mitchell Hashimoto, Ghostty creator, embeds prompt injections in codebases to detect and ban contributors submitting unreviewed AI-generated codeThe traps are placed in files like AGENTS.md.

16h|383.1k4.5k1k

MH#1562|@MITCHELLH

MM#1452|@PHOTOMATT

230

Gabriel Petersson, who worked on OpenAI's Sora, confirms he left the company a month before the official announcementAn outside user leaked his departure weeks before the announcement.

21h|1.5M15.4k877

GA#598|@GABRIEL1

RS#447|@SCOBLEIZER

240

Stripe's Patrick Collison says the Tempo blockchain hit a $3 billion annualized payment volume run rate 93 days after launchThe network hosts 1,000 payment services for AI agents.

14h|149.3k653244

MH#1108|@MATTHUANG

PC#81|@PATRICKC

250

Study of 26,811 students finds generative AI cuts homework time but reduces exam scores by 20%AI hurts learning when it replaces active mental effort.

15h|62.1k769321

XL#1855|@XLR8HARDER

EM#184|@EMOLLICK

260

Pirate Wires report says Medium burned up to $2.6 million a month while losing writers to SubstackMedium raised $157 million before tightening content moderation

15h|107.9k608184

MS#492|@MICSOLANA

RP#395|@TYPESFAST

270

Noah Smith parodies formulaic tech career announcements, mocking rapid shifts between hyped sectorsAntonio García Martínez called these updates an insufferable trope.

13h|152.8k1.2k67

AG#690|@ANTONIOGM

280

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

1d|76.6k69560

RP#1257|@ROHANPAUL_AI

PT#821|@ELDER_PLINIUS

B(#530|@BEFFJEZOS

290

Audit of Google DeepMind's DiffusionGemma finds intermediate steps remain human-interpretable, enabling chain-of-thought safety monitoringThis restores transparency benefits previously restricted to autoregressive LLMs.

12h|15.9k212108

ST#1694|@_ONIONESQUE

AC#1278|@ARTHURCONMY

BO#831|@BODONOGHUE85

NN#254|@NEELNANDA5

300

Systems engineer Yacine argues tech companies should stop hiring from NeurIPS and CVPR and recruit from SIGGRAPH insteadHeinrich Kuttler noted DeepMind's original team relied on gaming.

12h|141.5k2.6k340

HE#708|@HEINRICHKUTTLER

KA#403|@YACINEMTB