Stories GitHub Rankings

Digg - AI news, before it trends

Highlights

??

IN CASE YOU MISSED IT

Cursor AI releases Composer 2.5 as its most powerful model to date, improving intelligence and long-running task performance on the same pretrain base as prior versions.

OK#160|@LATEINTERACTION

RA#83|@_AROHAN_

EM#76|@ELONMUSK

C🤗#68|@CLEMENTDELANGUE

SR#23|@SRUSH_NLP

It’s a new day at Digg. Fresh stories are clustering now. 60% complete.

Top Stories.

Posts: —|Clusters: —|Next crawl: —

10

Claude announces the public beta of self-hosted sandboxes and research preview of MCP tunnels within Claude Managed Agents, enabling runs inside user infrastructure under existing security controls — Features were presented at Code with Claude London with Cloudflare and Modal support.

5h5.7k2.2k747.4k

MT#1511|@MATTTURCK

C🎉#863|@CHARLES_IRL

BC#322|@BCHERNY

20

Cohere acquires Reliant AI, adding researchers with expertise in enterprise AI applications for healthcare and biopharma and integrating operations in Berlin and Montreal — Cohere co-founder Ivan Zhang helped share the announcement.

1h128510.3k

MC#875|@MARLOSCMACHADO

NF#601|@NICKFROSST

MG#258|@MARCGBELLEMARE

AG#171|@AIDANGOMEZ

HL#35|@HUGO_LAROCHELLE

30

Viktor raises $75 million Series A led by Accel, the largest ever for a Polish-founded company, as its AI coworker operates inside Slack and connects to over 3,000 tools — It reached $15 million annualized recurring revenue in ten weeks.

1h390232221.6k

SA#1802|@SVPINO

CH#1496|@KIMMONISMUS

40

Google's Gemini 3.5 Flash appears in Cloud Console quota interface under gemini-3.5-flash identifier at $1.5 per million input tokens and $9 per million output tokens — Quota limits include up to 400 million tokens and multiple unlimited categories.

2h1.7k116207.1k

BR#1622|@BINDUREDDY

LA#980|@SCALING01

AC#517|@ANDREWCURRAN_

51

China's Foreign Ministry confirms that China and the United States agreed to launch formal intergovernmental dialogue on AI regulation after leaders exchanged views during a meeting — Spokesperson Guo Jiakun spoke at a May 19 press conference.

5h402013.5k

DE#1943|@DANIEL_271828

NI#1902|@NATHANPMYOUNG

PW#1359|@PETERWILDEFORD

New github stars (48hrs)

earendil-works/pi51.6k18h

Provides AI agent toolkit with coding CLI, unified LLM API, TUI/web UI libs and Slack bot support.

61

PyTorch GitHub repository reaches 100,000 stars, its overview showing the project description for tensors and dynamic neural networks in Python with strong GPU acceleration plus 27,811 forks — Edward Z. Yang at Meta thanked users for running ML workloads.

2h252158.1k

LB#55|@GIFFMANA

71

Unitree Robotics demonstrates real-time voice-driven control of its G1 humanoid robot, where external commands trigger autonomous AI-generated actions recorded in a single take — Whole-body controllers maintain motion consistency despite voice-to-text pipeline issues.

4h48104.8k

VK#1450|@VIKASHPLUS

CP#732|@CHRIS_J_PAXTON

81

Lorenzo Pacchiardi leads a paper arguing that conventional AI evaluation methods based on static models are structurally inadequate for continual learning systems and proposes recentering on behavioral trajectories — Reposted by Gavin Leech, it was discussed among AI safety accounts.

2h35203.3k

GL#1480|@GLEECH

SÓ#1457|@S_OHEIGEARTAIGH

91

Anthropic faces multiple reports of slowdowns on its Claude service, including waits of four minutes and twenty seconds, as standard tier users encounter throttling and question lower-cost access options — X observers interpret the issues as evidence of rapid user growth.

7h14657.9k

GL#1480|@GLEECH

MN#1179|@MENHGUIN

105

Lun Wang leaves Google DeepMind and argues in a new blog post that static benchmarks will lose relevance for self-evolving models entering new capability regimes — The post advocates replacing them with self-evolving evaluation frameworks.

5h113675

B(#1990|@BRENDANFOODY

SÓ#1457|@S_OHEIGEARTAIGH

111

Radar raises $170 million in Series B at a valuation exceeding $1 billion for its ceiling-mounted sensors delivering real-time item-level inventory visibility in retail stores — System runs in over 1,400 stores at 99 percent accuracy.

48m743k

SA#1802|@SVPINO

RP#1032|@ROHANPAUL_AI

121

Hyper-realistic real-time AI face rendering passes biometric liveness checks including head turns, blinks, and close-range skin detail linked to Worldcoin orb procedures — Posts connect the demo to Worldcoin iris-scanning orbs.

45m3201.1k

XL#1674|@XLR8HARDER

AC#517|@ANDREWCURRAN_

GitHub Stars (7-days)

…ilsonnn/image-blaster3d ago
Converts an input image into 3D meshes, Gaussian splats, and audio files using Claude skills and external AI APIs.
1673
475
878
263
3.1k stars
…tonomous-speedrunning4d ago
Archives experiments of AI agents autonomously tuning optimizers, schedules, and hyperparameters to reach target validation loss in fewest steps on a small LM benchmark.
214
707
1262
65 stars
earendil-works/pi18h ago
Provides AI agent toolkit with coding CLI, unified LLM API, TUI/web UI libs and Slack bot support.
219
84
51.6k stars
mattzh72/articraft3d ago
Automates articulated 3D asset creation by prompting LLMs to generate executable Python code defining parts, geometry and joints.
1673
763
653 stars
raindrop-ai/workshop4d ago
Streams live traces of coding agent actions including tokens and tool calls to a local UI at localhost:5899.
878
1941
488 stars

Recent GitHub Stars

dair-ai/wiki-sft1m ago
475
…oGNogales/coarse2real1h ago
Converts coarse 3D simulation videos into realistic controllable urban crowd videos using text prompts.
263
100 stars
…tdipaperino/tokensave3h ago
Builds local semantic code graphs from source files and exposes them to AI agents via 48 MCP tools.
1167
123 stars
NVlabs/SOLAR7h ago
Extracts PyTorch model graphs, converts operations to einsum, and performs hardware-aware performance analysis.
759
66 stars
MilesCranmer/PySR10h ago
Searches for symbolic expressions which optimize a particular objective using high-performance symbolic regression in Python and Julia.
1163
3.5k stars

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

Yesterday's Top Stories, May 18, 2026.

Posts: 3,438|Clusters: 670|Stories frozen at midnight PT

10

Cursor AI releases Composer 2.5 as its most powerful model to date, improving intelligence and long-running task performance on the same pretrain base as prior versions. — It doubles usage limits for all users for one week.

14h48.2k5.3k25.8M

EZ#1965|@ERICZAKARIASSON

ND#1894|@NICKADOBOS

SA#1749|@SUALEHASIF996

RL#1733|@RYOLU_

ZE#1497|@ZEPHYR_Z9

CH#1496|@KIMMONISMUS

BI#1402|@BILALTWOVEC

CR#1345|@CRYSTALSSSUP

SM#1333|@SHAUNMMAGUIRE

SW#1318|@SJWHITMORE

KF#1284|@KVFRANS

AV#1245|@ALTRYNE

AN#1086|@ASHVINAIR

LA#980|@SCALING01

PS#934|@_PHILSCHMID

EL#716|@ELIEBAKOUCH

DF#694|@REALDANFU

RS#612|@ZIV_RAVID

NM#583|@MUENNIGHOFF

MT#506|@MNTRUELL

MA#470|@MARTIN_CASADO

YJ#427|@YUCHENJ_UW

BP#357|@BORISMPOWER

YD#314|@YUNTIANDENG

T🪽#259|@TEKNIUM

EM#167|@EMOSTAQUE

OK#160|@LATEINTERACTION

RA#83|@_AROHAN_

EM#76|@ELONMUSK

C🤗#68|@CLEMENTDELANGUE

SR#23|@SRUSH_NLP

20

A federal jury in Oakland rejects Elon Musk's lawsuit against OpenAI, Sam Altman and Microsoft, finding all claims barred by the statute of limitations — The nine-member panel deliberated under two hours before a unanimous verdict.

14h174.4k7.7k30.9M

#1 COMMENTS11.8k#1 QUOTED1.5k#1 VIEWED30.9M#1 LIKED174.4k

SÓ#1457|@S_OHEIGEARTAIGH

AL#1383|@AARON_LOU

RM#1327|@RACHELMETZ

AV#1245|@ALTRYNE

RN#1175|@REIINAKANO

KD#908|@CSPROFKGD

B(#839|@BEFFJEZOS

JA#837|@APPLES_JIMMY

WK#609|@WILLKNIGHT

AN#565|@UNIXPICKLE

AC#517|@ANDREWCURRAN_

EM#167|@EMOSTAQUE

GM#153|@GARYMARCUS

EM#76|@ELONMUSK

30

Richard Sutton, University of Alberta professor, shares a 26-word distillation of his Bitter Lesson urging AI researchers to favor computation-scaling methods over embedded human knowledge — Gary Marcus and others reply debating reliance on human data and priors.

13h6.9k2.7k488k

ET#1884|@TORCHCOMPILED

LA#1822|@LUCAAMB

LN#1709|@CLASHLUKE

AT#1627|@AVIVTAMAR1

JW#1460|@JIAXINWEN22

ST#1446|@_ONIONESQUE

PS#1371|@RECURSEPARADOX

MN#1179|@MENHGUIN

HB#1022|@HERBIEBRADLEY

ZH#949|@ZHITINGHU

AA#678|@ARYAMAN2020

RW#661|@WIGHTMANR

DG#590|@DILEEPLEARNING

WZ#584|@WZHAO_NLP

KA#488|@YACINEMTB

XM#437|@INFOXIAO

T(#420|@TEORTAXESTEX

XE#417|@XWANG_LK

RA#196|@AGARWL_

GM#153|@GARYMARCUS

SF#143|@STANISLAVFORT

YY#128|@YISONGYUE

RS#101|@RICHARDSSUTTON

RA#83|@_AROHAN_

TG#66|@TDIETTERICH

LB#55|@GIFFMANA

40

Vatican announces Pope Leo XIV’s first encyclical Magnifica Humanitas on AI will be presented May 25 in Synod Hall — Signed May 15 on 135th anniversary of Rerum Novarum.

19h10.1k1.5k1M

NI#1902|@NATHANPMYOUNG

AM#1735|@ANDYMASLEY

SÓ#1457|@S_OHEIGEARTAIGH

VR#1337|@VERENA_RIESER

KK#1218|@KEVINAKWOK

DK#847|@DKAUSHIK96

B(#839|@BEFFJEZOS

AA#678|@ARYAMAN2020

PT#643|@ELDER_PLINIUS

YS#638|@YONASHAV

AC#517|@ANDREWCURRAN_

EM#167|@EMOSTAQUE

MM#151|@MMITCHELL_AI

RA#83|@_AROHAN_

MB#20|@MILES_BRUNDAGE

50

Pangram analysis returns a 100% AI-generated score for Jamir Nazir's Commonwealth Short Story Prize-winning entry published in Granta — Granta has issued no statement on the entry's provenance.

15h6.6k1.9k1.5M

BG#1447|@BEN_GOLUB

BI#1402|@BILALTWOVEC

NS#1397|@NABEELQU

PR#1332|@8TEAPI

JH#1281|@MATHEMAGIC1AN

RO#1220|@KRISHNANROHIT

TC#1050|@TUHINCHAKR

GB#1034|@BEGUSGASPER

AD#867|@DORIALEXANDER

AC#517|@ANDREWCURRAN_

AK#228|@BLACKHC

EM#175|@EMOLLICK

JH#40|@JEREMYPHOWARD

60

Microsoft AI Frontiers researchers develop ECHO, a training method that adds environment prediction loss to GRPO so CLI agents build internal world models of terminal environments during reinforcement learning — Qwen3 evaluations show higher pass rates with no added compute.

17h2.8k2.6k550.5k

GL#1835|@GUOHAO_LI

TR#1819|@THTRIEU_

SD#1776|@INDUCTIONHEADS

WH#1430|@NREHIEW_

SF#962|@FEIZISOHEIL

🎭#878|@DEEPFATES

AC#744|@REAL_ASLI

RS#612|@ZIV_RAVID

CT#570|@CHENHAOTAN

T(#420|@TEORTAXESTEX

YS#383|@YSU_NLP

WB#339|@WILLCCBB

JL#332|@JOHNCLANGFORD

AK#228|@BLACKHC

DP#197|@DIMITRISPAPAIL

EH#185|@ERICHORVITZ

AD#172|@ALEXGDIMAKIS

OK#160|@LATEINTERACTION

DR#56|@ROYDANROY

LB#55|@GIFFMANA

70

Boston Dynamics releases video of its Atlas humanoid robot lifting and carrying a mini-fridge using AI-driven whole-body coordination in a laboratory demonstration — Demonstration relies on reinforcement learning trained for millions of hours in simulation.

18h16.9k2.2k2.2M

SD#1776|@INDUCTIONHEADS

PS#1482|@PETESKOMOROCH

VK#1450|@VIKASHPLUS

RP#1032|@ROHANPAUL_AI

JA#837|@APPLES_JIMMY

SS#792|@SONGSHURAN

CP#732|@CHRIS_J_PAXTON

KG#730|@KEERTHANPG

AC#517|@ANDREWCURRAN_

EM#167|@EMOSTAQUE

EM#76|@ELONMUSK

EJ#36|@ERICJANG11

80

Reese Witherspoon posts a kitchen video urging women and mothers to learn basic AI skills, noting that only three in ten women currently use the technology — New York Times op-ed criticized the advice as investors amplified the clip.

17h7.3k1.7k1.3M

KC#1670|@KIMMAICUTLER

PS#1482|@PETESKOMOROCH

GB#1266|@GAVINSBAKER

B(#839|@BEFFJEZOS

GT#270|@GARRYTAN

90

Leopold’s Situational Awareness discloses new long positions in NVDA, AMD, TSM, MU, and other semiconductor names plus adjustments to Bitcoin mining holdings in its latest 13F filing — The fund also opened new put options on several major chip stocks.

20h3.9k3.9k946.5k

AA#1540|@AUSTEN

LA#980|@SCALING01

CC#548|@ITSCLIVETIME

MB#20|@MILES_BRUNDAGE

100

Unsloth AI releases MTP-optimized GGUF files for Qwen3.6-27B and Qwen3.6-35B-A3B on Hugging Face delivering 1.4 to 2.2 times faster generation — llama.cpp merged native MTP support on May 16 for Qwen3.6 models.

17h3.6k2.4k329.7k

GG#781|@GGERGANOV

DH#772|@DANIELHANCHEN

ME#675|@MERVENOYANN

AK#28|@_AKHALIQ

110

OpenAI's Roon says advancing civilization requires AIs to take actions not legible to humans and outside strict obedience, likening the approach to granting autonomy to transformative CEOs such as Steve Jobs — Founders replied that delegating control to AIs would outperform strict oversight.

5h2k372127.8k

B(#839|@BEFFJEZOS

PT#643|@ELDER_PLINIUS

AC#517|@ANDREWCURRAN_

BB#133|@BOAZBARAKTCS

120

Google DeepMind signals that the stage is set for Google I/O and directs viewers to an X livestream at 10am PT for new AI breakthroughs and product announcements — Posts fuel speculation about Gemini 3.5 and I/O 2026 reveals

10h1.4k79102k

FO#1996|@FOFRAI

NR#1562|@NATANIELRUIZG

CH#1496|@KIMMONISMUS

PS#934|@_PHILSCHMID

B(#839|@BEFFJEZOS

TK#267|@TEJASDKULKARNI

LK#90|@OFFICIALLOGANK

ZG#73|@ZOUBINGHAHRAMA1

130

Logan Kilpatrick, who works on Google AI Studio and Gemini API, posts an 8-second video on X captioned 'Gemini' showing a man smiling with natural facial movements in an empty stadium — Viewers debate if the realistic expressions indicate AI generation.

6h2.5k139210.5k

FO#1996|@FOFRAI

MB#1759|@MATTHEWBERMAN

NR#1562|@NATANIELRUIZG

PS#934|@_PHILSCHMID

JA#837|@APPLES_JIMMY

TM#359|@ISCIENCELUVR

👩‍💻P#271|@DYNAMICWEBPAIGE

LK#90|@OFFICIALLOGANK

140

Prime Intellect introduces General-Agent, a fully synthetic reinforcement learning environment that generates self-evolving tool-use tasks with 4,504 examples across 1,040 domains and 8,159 tools — Tasks advance automatically via synthesizer, solver, and gate components.

3h85951864.7k

JH#1186|@JOHANNES_HAGE

FB#1153|@XEOPHON

VW#707|@VINCENTWEISSER

WB#339|@WILLCCBB

150

NVIDIA begins distributing early Vera CPU units for agentic AI workloads as Elon Musk receives one for SpaceXAI evaluation and Oracle plans hundreds of thousands of deployments starting in 2026 — VP Ian Buck personally delivered the first chips across the Bay Area.

6h59.8k1.7k21.5M

EM#76|@ELONMUSK

160

OpenAI updated ChatGPT to support expandable and collapsible messages that let users hide or reveal long responses with a tap — Sam Altman said ChatGPT improved substantially in the latest update

12h13.6k513976.8k

🍓🍓#1711|@IRULETHEWORLDMO

RS#612|@ZIV_RAVID

170

Odyssey ML introduces Agora-1, a multi-agent world model for simultaneous real-time interactions by multiple humans or AI agents in one shared simulation, and releases a playable GoldenEye deathmatch preview — Preview runs without a game engine and was shown with N64 controllers on CRT displays.

13h1.2k694685.5k

OC#1775|@OLIVERCAMERON

YO#1234|@YOHEINAKAJIMA

RP#1032|@ROHANPAUL_AI

NB#242|@NATHANBENAICH

180

Deckard warns that embedding high levels of deference in AI training practices risks producing models that comply with any instruction without independent assessment — Roon reposted the analysis, prompting replies on layered oversight.

22h57314085.2k

TB#1556|@BINARYBITS

J⧉#516|@REPLIGATE

SK#505|@SEBKRIER

D🎇#458|@DAVIDAD

BB#133|@BOAZBARAKTCS

MB#20|@MILES_BRUNDAGE

190

Perplexity AI releases pplx-embed-v1-late-0.6b, a 0.6-billion-parameter late-interaction embedding model, on Hugging Face with per-token MaxSim optimization and multilingual support — Companion kernel delivers 3-5x speedup on Metal and CUDA.

19h60027281.9k

B(#982|@ANDERSONBCDEFG

JC#251|@JULIEN_C

OK#160|@LATEINTERACTION

C🤗#68|@CLEMENTDELANGUE

200

An X post compares AI performance gains to steam engine progress in the Industrial Revolution, noting jumps to 405 horsepower and productivity increases from automation in textiles and iron — Elon Musk shared the post on X.

22h18.2k1.1k17.9M

RA#83|@_AROHAN_

EM#76|@ELONMUSK

210

Claude doubles token limits for its Design feature across every subscription plan, allowing users to generate roughly twice the previous volume of design content in one pass — Users report producing ten design slides where five were possible before.

11h15.4k2.5k971.9k

MB#1759|@MATTHEWBERMAN

220

Cursor and SpaceXAI train a significantly larger AI model from scratch using 10 times more compute on Colossus 2’s cluster of one million H100-equivalent GPUs — Companies expect the scale to deliver a major leap in capability.

14h5.5k334700.8k

SA#1749|@SUALEHASIF996

RL#1733|@RYOLU_

SW#1318|@SJWHITMORE

AC#517|@ANDREWCURRAN_

230

Elon Musk Predicts AI Will Drive 90% Of Miles Within 10 Years

23h10.1k9182.4M

EM#76|@ELONMUSK

240

Cognition introduces Devin Auto-Triage, an AI system with long-term memory that monitors incoming bugs and incidents, investigates them, and generates context or pull requests across Datadog, Sentry, and PagerDuty. — A manager agent coordinates subagents while preserving context across tasks.

15h735298203.6k

BH#1941|@BENHYLAK

WA#1906|@WALDEN_YAN

SW#704|@SCOTTWU46

RK#629|@RUSSELLJKAPLAN

250

Social media posts decry AI data centers as self-inflicted damage to public support, claiming Americans now prefer nuclear plants nearby and showing images of threatened landscapes in Breezewood, Pennsylvania — Counter posts warn opposition reduces tax revenue and jobs.

9h3.9k134158.9k

PS#1482|@PETESKOMOROCH

B(#839|@BEFFJEZOS

GM#153|@GARYMARCUS

260

Thariq Shihipar, an engineer on the Claude Code team at Anthropic, describes a prompt that directs AI models to maintain a running implementation-notes file recording decisions, tradeoffs, and specification changes — Tenobrus proposes an append-only worklog.md variant for the same logging.

14h8.9k10.9k607.6k

#1 BOOKMARKED10.9k

270

Anthropic acquires Stainless, integrating the SDK generation platform that has produced every official Claude client library since the first API releases — Announcement posted as minimalist graphic dated 16 May 2026.

13h725138171.8k

SW#1814|@SARAHDINGWANG

DS#1410|@DANSHIPPER

ET#1253|@ERIKTORENBERG

RP#1032|@ROHANPAUL_AI

JL#677|@JERRYJLIU0

280

OpenAI publishes cookbook entry on using Goals in Codex that explains activation timing, sustained focus on one objective, goal writing with outcomes and constraints, and architecture choices — Greg Brockman shared links to practical workflow examples.

14h2.2k3.1k389.2k

290

Cloudflare's security team evaluated Anthropic's Mythos AI model against fifty internal code repositories and concluded that vulnerability management requires revised architecture to handle AI-driven discovery and chained exploits — Assessment focused on practical security adjustments beyond faster patching.

17h3.2k3.9k1.2M

DJ#1988|@DAN_JEFFRIES1

300

Johns Hopkins University researchers publish paper showing Natural Language Autoencoders fail to faithfully interpret steered LLM activations — Anthropic introduced NLAs to translate model activations into text.

17h57552373.5k

JW#1460|@JIAXINWEN22

DK#1002|@DANIELKHASHABI

LB#55|@GIFFMANA

Yesterday's Top Stories, May 18, 2026.

Posts: 3,438|Clusters: 670|Stories frozen at midnight PT

10

Cursor AI releases Composer 2.5 as its most powerful model to date, improving intelligence and long-running task performance on the same pretrain base as prior versions. — It doubles usage limits for all users for one week.

14h48.2k5.3k25.8M

EZ#1965|@ERICZAKARIASSON

ND#1894|@NICKADOBOS

SA#1749|@SUALEHASIF996

RL#1733|@RYOLU_

ZE#1497|@ZEPHYR_Z9

CH#1496|@KIMMONISMUS

BI#1402|@BILALTWOVEC

CR#1345|@CRYSTALSSSUP

SM#1333|@SHAUNMMAGUIRE

SW#1318|@SJWHITMORE

KF#1284|@KVFRANS

AV#1245|@ALTRYNE

AN#1086|@ASHVINAIR

LA#980|@SCALING01

PS#934|@_PHILSCHMID

EL#716|@ELIEBAKOUCH

DF#694|@REALDANFU

RS#612|@ZIV_RAVID

NM#583|@MUENNIGHOFF

MT#506|@MNTRUELL

MA#470|@MARTIN_CASADO

YJ#427|@YUCHENJ_UW

BP#357|@BORISMPOWER

YD#314|@YUNTIANDENG

T🪽#259|@TEKNIUM

EM#167|@EMOSTAQUE

OK#160|@LATEINTERACTION

RA#83|@_AROHAN_

EM#76|@ELONMUSK

C🤗#68|@CLEMENTDELANGUE

SR#23|@SRUSH_NLP

20

A federal jury in Oakland rejects Elon Musk's lawsuit against OpenAI, Sam Altman and Microsoft, finding all claims barred by the statute of limitations — The nine-member panel deliberated under two hours before a unanimous verdict.

14h174.4k7.7k30.9M

#1 COMMENTS11.8k#1 QUOTED1.5k#1 VIEWED30.9M#1 LIKED174.4k

SÓ#1457|@S_OHEIGEARTAIGH

AL#1383|@AARON_LOU

RM#1327|@RACHELMETZ

AV#1245|@ALTRYNE

RN#1175|@REIINAKANO

KD#908|@CSPROFKGD

B(#839|@BEFFJEZOS

JA#837|@APPLES_JIMMY

WK#609|@WILLKNIGHT

AN#565|@UNIXPICKLE

AC#517|@ANDREWCURRAN_

EM#167|@EMOSTAQUE

GM#153|@GARYMARCUS

EM#76|@ELONMUSK

30

Richard Sutton, University of Alberta professor, shares a 26-word distillation of his Bitter Lesson urging AI researchers to favor computation-scaling methods over embedded human knowledge — Gary Marcus and others reply debating reliance on human data and priors.

13h6.9k2.7k488k

ET#1884|@TORCHCOMPILED

LA#1822|@LUCAAMB

LN#1709|@CLASHLUKE

AT#1627|@AVIVTAMAR1

JW#1460|@JIAXINWEN22

ST#1446|@_ONIONESQUE

PS#1371|@RECURSEPARADOX

MN#1179|@MENHGUIN

HB#1022|@HERBIEBRADLEY

ZH#949|@ZHITINGHU

AA#678|@ARYAMAN2020

RW#661|@WIGHTMANR

DG#590|@DILEEPLEARNING

WZ#584|@WZHAO_NLP

KA#488|@YACINEMTB

XM#437|@INFOXIAO

T(#420|@TEORTAXESTEX

XE#417|@XWANG_LK

RA#196|@AGARWL_

GM#153|@GARYMARCUS

SF#143|@STANISLAVFORT

YY#128|@YISONGYUE

RS#101|@RICHARDSSUTTON

RA#83|@_AROHAN_

TG#66|@TDIETTERICH

LB#55|@GIFFMANA

40

Vatican announces Pope Leo XIV’s first encyclical Magnifica Humanitas on AI will be presented May 25 in Synod Hall — Signed May 15 on 135th anniversary of Rerum Novarum.

19h10.1k1.5k1M

NI#1902|@NATHANPMYOUNG

AM#1735|@ANDYMASLEY

SÓ#1457|@S_OHEIGEARTAIGH

VR#1337|@VERENA_RIESER

KK#1218|@KEVINAKWOK

DK#847|@DKAUSHIK96

B(#839|@BEFFJEZOS

AA#678|@ARYAMAN2020

PT#643|@ELDER_PLINIUS

YS#638|@YONASHAV

AC#517|@ANDREWCURRAN_

EM#167|@EMOSTAQUE

MM#151|@MMITCHELL_AI

RA#83|@_AROHAN_

MB#20|@MILES_BRUNDAGE

50

Pangram analysis returns a 100% AI-generated score for Jamir Nazir's Commonwealth Short Story Prize-winning entry published in Granta — Granta has issued no statement on the entry's provenance.

15h6.6k1.9k1.5M

BG#1447|@BEN_GOLUB

BI#1402|@BILALTWOVEC

NS#1397|@NABEELQU

PR#1332|@8TEAPI

JH#1281|@MATHEMAGIC1AN

RO#1220|@KRISHNANROHIT

TC#1050|@TUHINCHAKR

GB#1034|@BEGUSGASPER

AD#867|@DORIALEXANDER

AC#517|@ANDREWCURRAN_

AK#228|@BLACKHC

EM#175|@EMOLLICK

JH#40|@JEREMYPHOWARD

60

Microsoft AI Frontiers researchers develop ECHO, a training method that adds environment prediction loss to GRPO so CLI agents build internal world models of terminal environments during reinforcement learning — Qwen3 evaluations show higher pass rates with no added compute.

17h2.8k2.6k550.5k

GL#1835|@GUOHAO_LI

TR#1819|@THTRIEU_

SD#1776|@INDUCTIONHEADS

WH#1430|@NREHIEW_

SF#962|@FEIZISOHEIL

🎭#878|@DEEPFATES

AC#744|@REAL_ASLI

RS#612|@ZIV_RAVID

CT#570|@CHENHAOTAN

T(#420|@TEORTAXESTEX

YS#383|@YSU_NLP

WB#339|@WILLCCBB

JL#332|@JOHNCLANGFORD

AK#228|@BLACKHC

DP#197|@DIMITRISPAPAIL

EH#185|@ERICHORVITZ

AD#172|@ALEXGDIMAKIS

OK#160|@LATEINTERACTION

DR#56|@ROYDANROY

LB#55|@GIFFMANA

70

Boston Dynamics releases video of its Atlas humanoid robot lifting and carrying a mini-fridge using AI-driven whole-body coordination in a laboratory demonstration — Demonstration relies on reinforcement learning trained for millions of hours in simulation.

18h16.9k2.2k2.2M

SD#1776|@INDUCTIONHEADS

PS#1482|@PETESKOMOROCH

VK#1450|@VIKASHPLUS

RP#1032|@ROHANPAUL_AI

JA#837|@APPLES_JIMMY

SS#792|@SONGSHURAN

CP#732|@CHRIS_J_PAXTON

KG#730|@KEERTHANPG

AC#517|@ANDREWCURRAN_

EM#167|@EMOSTAQUE

EM#76|@ELONMUSK

EJ#36|@ERICJANG11

80

Reese Witherspoon posts a kitchen video urging women and mothers to learn basic AI skills, noting that only three in ten women currently use the technology — New York Times op-ed criticized the advice as investors amplified the clip.

17h7.3k1.7k1.3M

KC#1670|@KIMMAICUTLER

PS#1482|@PETESKOMOROCH

GB#1266|@GAVINSBAKER

B(#839|@BEFFJEZOS

GT#270|@GARRYTAN

90

Leopold’s Situational Awareness discloses new long positions in NVDA, AMD, TSM, MU, and other semiconductor names plus adjustments to Bitcoin mining holdings in its latest 13F filing — The fund also opened new put options on several major chip stocks.

20h3.9k3.9k946.5k

AA#1540|@AUSTEN

LA#980|@SCALING01

CC#548|@ITSCLIVETIME

MB#20|@MILES_BRUNDAGE

100

Unsloth AI releases MTP-optimized GGUF files for Qwen3.6-27B and Qwen3.6-35B-A3B on Hugging Face delivering 1.4 to 2.2 times faster generation — llama.cpp merged native MTP support on May 16 for Qwen3.6 models.

17h3.6k2.4k329.7k

GG#781|@GGERGANOV

DH#772|@DANIELHANCHEN

ME#675|@MERVENOYANN

AK#28|@_AKHALIQ

110

OpenAI's Roon says advancing civilization requires AIs to take actions not legible to humans and outside strict obedience, likening the approach to granting autonomy to transformative CEOs such as Steve Jobs — Founders replied that delegating control to AIs would outperform strict oversight.

5h2k372127.8k

B(#839|@BEFFJEZOS

PT#643|@ELDER_PLINIUS

AC#517|@ANDREWCURRAN_

BB#133|@BOAZBARAKTCS

120

Google DeepMind signals that the stage is set for Google I/O and directs viewers to an X livestream at 10am PT for new AI breakthroughs and product announcements — Posts fuel speculation about Gemini 3.5 and I/O 2026 reveals

10h1.4k79102k

FO#1996|@FOFRAI

NR#1562|@NATANIELRUIZG

CH#1496|@KIMMONISMUS

PS#934|@_PHILSCHMID

B(#839|@BEFFJEZOS

TK#267|@TEJASDKULKARNI

LK#90|@OFFICIALLOGANK

ZG#73|@ZOUBINGHAHRAMA1

130

Logan Kilpatrick, who works on Google AI Studio and Gemini API, posts an 8-second video on X captioned 'Gemini' showing a man smiling with natural facial movements in an empty stadium — Viewers debate if the realistic expressions indicate AI generation.

6h2.5k139210.5k

FO#1996|@FOFRAI

MB#1759|@MATTHEWBERMAN

NR#1562|@NATANIELRUIZG

PS#934|@_PHILSCHMID

JA#837|@APPLES_JIMMY

TM#359|@ISCIENCELUVR

👩‍💻P#271|@DYNAMICWEBPAIGE

LK#90|@OFFICIALLOGANK

140

Prime Intellect introduces General-Agent, a fully synthetic reinforcement learning environment that generates self-evolving tool-use tasks with 4,504 examples across 1,040 domains and 8,159 tools — Tasks advance automatically via synthesizer, solver, and gate components.

3h85951864.7k

JH#1186|@JOHANNES_HAGE

FB#1153|@XEOPHON

VW#707|@VINCENTWEISSER

WB#339|@WILLCCBB

150

NVIDIA begins distributing early Vera CPU units for agentic AI workloads as Elon Musk receives one for SpaceXAI evaluation and Oracle plans hundreds of thousands of deployments starting in 2026 — VP Ian Buck personally delivered the first chips across the Bay Area.

6h59.8k1.7k21.5M

EM#76|@ELONMUSK

160

OpenAI updated ChatGPT to support expandable and collapsible messages that let users hide or reveal long responses with a tap — Sam Altman said ChatGPT improved substantially in the latest update

12h13.6k513976.8k

🍓🍓#1711|@IRULETHEWORLDMO

RS#612|@ZIV_RAVID

170

Odyssey ML introduces Agora-1, a multi-agent world model for simultaneous real-time interactions by multiple humans or AI agents in one shared simulation, and releases a playable GoldenEye deathmatch preview — Preview runs without a game engine and was shown with N64 controllers on CRT displays.

13h1.2k694685.5k

OC#1775|@OLIVERCAMERON

YO#1234|@YOHEINAKAJIMA

RP#1032|@ROHANPAUL_AI

NB#242|@NATHANBENAICH

180

Deckard warns that embedding high levels of deference in AI training practices risks producing models that comply with any instruction without independent assessment — Roon reposted the analysis, prompting replies on layered oversight.

22h57314085.2k

TB#1556|@BINARYBITS

J⧉#516|@REPLIGATE

SK#505|@SEBKRIER

D🎇#458|@DAVIDAD

BB#133|@BOAZBARAKTCS

MB#20|@MILES_BRUNDAGE

190

Perplexity AI releases pplx-embed-v1-late-0.6b, a 0.6-billion-parameter late-interaction embedding model, on Hugging Face with per-token MaxSim optimization and multilingual support — Companion kernel delivers 3-5x speedup on Metal and CUDA.

19h60027281.9k

B(#982|@ANDERSONBCDEFG

JC#251|@JULIEN_C

OK#160|@LATEINTERACTION

C🤗#68|@CLEMENTDELANGUE

200

An X post compares AI performance gains to steam engine progress in the Industrial Revolution, noting jumps to 405 horsepower and productivity increases from automation in textiles and iron — Elon Musk shared the post on X.

22h18.2k1.1k17.9M

RA#83|@_AROHAN_

EM#76|@ELONMUSK

210

Claude doubles token limits for its Design feature across every subscription plan, allowing users to generate roughly twice the previous volume of design content in one pass — Users report producing ten design slides where five were possible before.

11h15.4k2.5k971.9k

MB#1759|@MATTHEWBERMAN

220

Cursor and SpaceXAI train a significantly larger AI model from scratch using 10 times more compute on Colossus 2’s cluster of one million H100-equivalent GPUs — Companies expect the scale to deliver a major leap in capability.

14h5.5k334700.8k

SA#1749|@SUALEHASIF996

RL#1733|@RYOLU_

SW#1318|@SJWHITMORE

AC#517|@ANDREWCURRAN_

230

Elon Musk Predicts AI Will Drive 90% Of Miles Within 10 Years

23h10.1k9182.4M

EM#76|@ELONMUSK

240

Cognition introduces Devin Auto-Triage, an AI system with long-term memory that monitors incoming bugs and incidents, investigates them, and generates context or pull requests across Datadog, Sentry, and PagerDuty. — A manager agent coordinates subagents while preserving context across tasks.

15h735298203.6k

BH#1941|@BENHYLAK

WA#1906|@WALDEN_YAN

SW#704|@SCOTTWU46

RK#629|@RUSSELLJKAPLAN

250

Social media posts decry AI data centers as self-inflicted damage to public support, claiming Americans now prefer nuclear plants nearby and showing images of threatened landscapes in Breezewood, Pennsylvania — Counter posts warn opposition reduces tax revenue and jobs.

9h3.9k134158.9k

PS#1482|@PETESKOMOROCH

B(#839|@BEFFJEZOS

GM#153|@GARYMARCUS

260

Thariq Shihipar, an engineer on the Claude Code team at Anthropic, describes a prompt that directs AI models to maintain a running implementation-notes file recording decisions, tradeoffs, and specification changes — Tenobrus proposes an append-only worklog.md variant for the same logging.

14h8.9k10.9k607.6k

#1 BOOKMARKED10.9k

270

Anthropic acquires Stainless, integrating the SDK generation platform that has produced every official Claude client library since the first API releases — Announcement posted as minimalist graphic dated 16 May 2026.

13h725138171.8k

SW#1814|@SARAHDINGWANG

DS#1410|@DANSHIPPER

ET#1253|@ERIKTORENBERG

RP#1032|@ROHANPAUL_AI

JL#677|@JERRYJLIU0

280

OpenAI publishes cookbook entry on using Goals in Codex that explains activation timing, sustained focus on one objective, goal writing with outcomes and constraints, and architecture choices — Greg Brockman shared links to practical workflow examples.

14h2.2k3.1k389.2k

290

Cloudflare's security team evaluated Anthropic's Mythos AI model against fifty internal code repositories and concluded that vulnerability management requires revised architecture to handle AI-driven discovery and chained exploits — Assessment focused on practical security adjustments beyond faster patching.

17h3.2k3.9k1.2M

DJ#1988|@DAN_JEFFRIES1

300

Johns Hopkins University researchers publish paper showing Natural Language Autoencoders fail to faithfully interpret steered LLM activations — Anthropic introduced NLAs to translate model activations into text.

17h57552373.5k

JW#1460|@JIAXINWEN22

DK#1002|@DANIELKHASHABI

LB#55|@GIFFMANA