Tech News

Today's Highlights

In case you missed it

Anthropic silently restricts Claude Fable 5 performance when detecting frontier LLM development tasks

NL#69|@NATOLAMBERT

DR#60|@ROYDANROY

LB#53|@GIFFMANA

MB#16|@MILES_BRUNDAGE

#1 VIEWED

Anthropic's Dario Amodei proposes mandatory third-party safety testing for frontier AI models to manage exponential capability growth

B(#824|@BEFFJEZOS

EH#607|@EVANHUB

AC#535|@ANDREWCURRAN_

T(#440|@TEORTAXESTEX

DA#173|@DARIOAMODEI

#1 LIKED

14.7k

Google DeepMind releases DiffusionGemma, an experimental 26B open-weights text diffusion model that generates 256-token blocks in parallel

EL#510|@OMARSAR0

AC#275|@ANDREW_N_CARR

SP#110|@SUNDARPICHAI

SD#88|@SEDIELEM

AK#34|@_AKHALIQ

MOST NEGATIVE

Extropic founder Guillaume Verdon claims Anthropic's strict safety filters make Claude unusable for biotechnology and chemistry research

RS#327|@SCOBLEIZER

JC#267|@JULIEN_C

AK#241|@BLACKHC

RS#49|@RICHARDSOCHER

JH#42|@JEREMYPHOWARD

Top Stories

Posts:—|Clusters:—|Next Crawl:—

11

Anthropic's Dario Amodei proposes mandatory third-party safety testing for frontier AI models to manage exponential capability growthCritics, including Steven Sinofsky, label the proposal regulatory capture.

2h|#1VIEWS1.7M9.9k#1BOOKMARKED5.1k#1COMMENTS1.4k

SS#1976|@STEVESI

GH#1761|@GHADFIELD

MB#1734|@BAKKERMICHIEL

SG#1662|@SHARONGOLDMAN

GL#1593|@GLEECH

CH#1532|@KIMMONISMUS

PS#1461|@PETESKOMOROCH

PW#1432|@PETERWILDEFORD

RO#1210|@KRISHNANROHIT

AL#1176|@ALTH0U

RP#1102|@ROHANPAUL_AI

HB#1081|@HERBIEBRADLEY

DF#1002|@DANIELLEFONG

RA#920|@0INTERESTRATES

SH#910|@SAFFRONHUANG

JA#900|@APPLES_JIMMY

B(#824|@BEFFJEZOS

EH#607|@EVANHUB

AC#535|@ANDREWCURRAN_

T(#440|@TEORTAXESTEX

DA#173|@DARIOAMODEI

21

Anthropic quietly restricts Claude Fable 5 from helping build ML accelerators and pretraining pipelinesThe silent limits affect roughly 0.03 percent of traffic.

14h|248.2k1.7k163

SS#1976|@STEVESI

SC#1848|@APSARATHCHANDAR

AM#1618|@ANDREWMAYNE

SH#1538|@HAMANDCHEESE

DS#1462|@DANSHIPPER

PS#1461|@PETESKOMOROCH

SH#1427|@SHYAMALANADKAT

SA#1383|@SAMSJA19

RP#1102|@ROHANPAUL_AI

LA#1064|@SCALING01

DF#1002|@DANIELLEFONG

AM#901|@ANDREWDAI

B(#824|@BEFFJEZOS

VW#743|@VINCENTWEISSER

RS#663|@ZIV_RAVID

ZN#589|@ZACHARYNADO

TR#514|@TALIARINGER

EL#510|@OMARSAR0

HP#488|@HYHIEU226

MG#484|@NLPMATTG

D🎇#469|@DAVIDAD

T(#440|@TEORTAXESTEX

DW#390|@DEANWBALL

ML#375|@MISHALASKIN

IO#365|@IANOSBAND

AK#241|@BLACKHC

SA#121|@PRFSANJEEVARORA

NL#70|@NATOLAMBERT

30

SemiAnalysis reports Anthropic's latest model filters and degrades machine learning research queries to prevent competitive or self-improving AI developmentCommentators warn these filters could trigger silent model sabotage.

15h|781.7k4.2k1.3k

CE#1686|@INTUITMACHINE

AP#1477|@PANDAASHWINEE

GO#1443|@GERGELYOROSZ

DF#1002|@DANIELLEFONG

AM#901|@ANDREWDAI

DF#753|@REALDANFU

CP#716|@CHAMATH

PH#676|@PETERHNDRSN

RT#640|@ROSSTAYLOR90

EC#550|@ETHANCABALLERO

EL#510|@OMARSAR0

🇺🇦D#405|@DBAHDANAU

RS#327|@SCOBLEIZER

RA#205|@AGARWL_

DR#111|@DELIPRAO

C🤗#74|@CLEMENTDELANGUE

Watchlist

Video Signals

QB#1726|@QBERTHET

CE#1686|@INTUITMACHINE

FB#1190|@XEOPHON

PV#601|@PETARV_93

AC#275|@ANDREW_N_CARR

SD#88|@SEDIELEM

Google Gemma Releases DiffusionGemma Open Model For Parallel Text Generation

7h ago|Views 309KLikes 3.3KBookmarks 1.3K

LA#1064|@SCALING01

CH#930|@CHATGPTAPP

LK#687|@LUKASZKAISER

ChatGPT Promotes MessiMode For Country Flag Hair Color Edits

7h ago|Views 1.4MLikes 9.4KBookmarks 1.1K

TR#514|@TALIARINGER

Anthropic Unveils Claude Fable 5 Mythos-Class Model for General Use

1d ago|Views 2.5MLikes 14KBookmarks 1.7K

BZ#1261|@BANGHUAZ

HP#488|@HYHIEU226

RS#327|@SCOBLEIZER

GY#114|@THEGREGYANG

River AI Launches to Build User-Owned Personal AI Systems

6h ago|Views 370KLikes 839Bookmarks 384

“P#1754|@PAULARAMBLES

Fable AI Autonomously Edits Its Own Launch Video With Code And Tools

1h ago|Views 95KLikes 1.9KBookmarks 2.6K

SW#1999|@SARAHDINGWANG

MM#598|@MASCOBOT

Cursor AI Grows From 15 To 700 Employees As 60 Percent Of Fortune 500 Adopt Platform

5h ago|Views 303KLikes 4KBookmarks 784

HC#780|@HWCHASE17

ML#375|@MISHALASKIN

Odessia Raises $6M Led By Sequoia For AI Travel Planning App

6h ago|Views 349KLikes 1.3KBookmarks 1.5K

CP#1119|@CHRISPAINTERYUP

NS#470|@NOAHPINION

DP#67|@DWARKESH_SP

Karpathy Predicts AGI Impact Will Blend Into Existing Economic Growth Rates

6h ago|Views 85KLikes 468Bookmarks 302

MO|@MOLLYSOSHEA

SpaceX Reaches $1.25T Valuation After Record xAI Acquisition

Mar 22|Views 1.4MLikes 2.9KBookmarks 445

MN#1274|@MENHGUIN

Claude Fable 5 Generates 9:16 Video From LLM Perspective

1d ago|Views 247KLikes 1.7KBookmarks 833

40

Google DeepMind releases DiffusionGemma, an experimental 26B open-weights text diffusion model that generates 256-token blocks in parallelIt runs locally within a 24GB VRAM envelope.

6h|1.1M#1LIKED14.7k4.4k

QB#1726|@QBERTHET

NR#1699|@NATANIELRUIZG

CH#1532|@KIMMONISMUS

OB#1314|@OLIVIERBACHEM

FB#1190|@XEOPHON

RP#1102|@ROHANPAUL_AI

KK#1095|@KASTNERKYLE

PS#1011|@_PHILSCHMID

DH#823|@DANIELHANCHEN

BO#755|@BODONOGHUE85

ME#722|@MERVENOYANN

PV#601|@PETARV_93

OS#511|@OSANSEVIERO

EL#510|@OMARSAR0

AC#275|@ANDREW_N_CARR

SP#110|@SUNDARPICHAI

SD#88|@SEDIELEM

AK#34|@_AKHALIQ

51

Malware developers bypass LLM security scanners by embedding biological and nuclear weapon reference strings to trigger safety refusalsThe packages target bioinformatics and Model Context Protocol developers

10h|1.1M9.9k3.2k

SS#1976|@STEVESI

ND#1923|@NICKADOBOS

GL#1593|@GLEECH

ZE#1561|@ZEPHYR_Z9

FB#1190|@XEOPHON

DF#1002|@DANIELLEFONG

B(#824|@BEFFJEZOS

PH#676|@PETERHNDRSN

SK#519|@SEBKRIER

TR#514|@TALIARINGER

KA#487|@YACINEMTB

New github stars (48hrs)

vllm-project/vime1791d

Integrates vLLM rollout with Megatron training for LLM post-training and RL scaling.

rdi-berkeley/agents-last-exam5211d

Provisions OS sandboxes, runs agent harnesses on long-horizon tasks, and grades outputs against hidden references.

xzf-thu/Audio-Interaction2831d

Implements the first unified always-on audio interaction model for streaming and offline tasks like ASR, translation, and proactive responses.

kyutai-labs/kairos71d

Trains 6B LLMs on temporally ordered Common Crawl data from 2018-2025 to measure recency bias and enable continual learning studies.

roboflow/supervision43.5k1d

Supplies unified Detections API, annotators, trackers, and dataset converters for computer vision pipelines.

61

Extropic founder Guillaume Verdon claims Anthropic's strict safety filters make Claude unusable for biotechnology and chemistry researchThe blocks reportedly impact both Claude and Fable 5.

15h|119.2k2.5k148

MOST NEGATIVE139

NI#2000|@NATHANPMYOUNG

SD#1958|@INDUCTIONHEADS

ND#1923|@NICKADOBOS

XL#1840|@XLR8HARDER

🍓🍓#1839|@IRULETHEWORLDMO

CE#1686|@INTUITMACHINE

TA#1096|@VICTORTAELIN

DF#1002|@DANIELLEFONG

B(#824|@BEFFJEZOS

ZN#589|@ZACHARYNADO

TR#514|@TALIARINGER

EL#510|@OMARSAR0

FF#352|@FRANCOISFLEURET

RS#327|@SCOBLEIZER

JC#267|@JULIEN_C

AK#241|@BLACKHC

RS#49|@RICHARDSOCHER

JH#42|@JEREMYPHOWARD

70

UK AI Security Institute Chief Scientist Geoffrey Irving launches Sequent to focus on high-confidence AI alignment and research automationThe organization plans to heavily automate its empirical research.

7h|137.9k1.8k430

NI#2000|@NATHANPMYOUNG

JP#1829|@JACOB_PFAU

SÓ#1606|@S_OHEIGEARTAIGH

GL#1593|@GLEECH

LI#1582|@LIVGORTON

JM#1517|@JESSEMHAN

NC#1435|@CHOWDHURYNEIL

TK#1271|@TOMEKKORBAK

SL#1178|@SETHLAZAR

RW#1153|@ROBERTWIBLIN

DK#620|@DKOKOTAJLO

GI#371|@GEOFFREYIRVING

EY#289|@ESYUDKOWSKY

SB#218|@BLANCHEMINERVA

DK#214|@DAVIDSKRUEGER

EP#161|@ETHANJPEREZ

81

Igor Babuschkin launches River AI to build a decentralized, user-owned personal AI stackThe startup plans to release its first products soon

6h|574k1.8k719

“P#1754|@PAULARAMBLES

VC#1319|@VITALIYCHILEY

BZ#1261|@BANGHUAZ

HP#488|@HYHIEU226

FI#415|@FINBARRTIMBERS

RS#327|@SCOBLEIZER

GY#114|@THEGREGYANG

91

Geth lead developer Péter Szilágyi argues Anthropic's Fable model restrictions on biology and cryptography represent corporate gatekeepingHe warns these safety limits sharply reduce the model's utility.

11h|556.8k11.8k1.1k

DF#1002|@DANIELLEFONG

B(#824|@BEFFJEZOS

VW#743|@VINCENTWEISSER

TR#514|@TALIARINGER

101

Sam Altman tells OpenAI staff an IPO is planned next year, but recursive self-improvement would favor staying privateThe memo also teased an upcoming model codenamed 5.6.

7h|341.1k1.5k335

SH#1538|@HAMANDCHEESE

CH#1532|@KIMMONISMUS

SP#1240|@STEPH_PALAZZOLO

RP#1102|@ROHANPAUL_AI

LA#1064|@SCALING01

ZN#589|@ZACHARYNADO

AC#535|@ANDREWCURRAN_

113

Former xAI co-founder Igor Babuschkin launches River AI, a startup building user-owned personal AI systemsThe startup plans to ship its initial stack soon.

6h|124.8k1.1k297

BA#1370|@BAYESLORD

B(#824|@BEFFJEZOS

EL#762|@ELIEBAKOUCH

EL#510|@OMARSAR0

AG#434|@ALEXGRAVELEY

RS#327|@SCOBLEIZER

EM#175|@EMOSTAQUE

121

Policy scholar Dean W. Ball argues Anthropic's safety policies constitute anticompetitive behavior disguised as AI safetyThe conduct weakens the argument for relaxing antitrust enforcement

11h|194.6k1.8k324

NI#2000|@NATHANPMYOUNG

PS#1461|@PETESKOMOROCH

DF#1002|@DANIELLEFONG

BT#570|@BENTHOMPSON

DW#390|@DEANWBALL

AK#241|@BLACKHC

BB#140|@BOAZBARAKTCS

132

White House AI advisor Sriram Krishnan warns that open science and singularity-focused AI acceleration are becoming incompatibleHe urged Western nations to preserve distributed compute and research.

13h|276.7k1.9k171

CF#1480|@CFGEEK

VW#743|@VINCENTWEISSER

EL#510|@OMARSAR0

SK#336|@SRIRAMK

GT#251|@GARRYTAN

143

The Pragmatic Engineer's Gergely Orosz criticizes Anthropic's Fable service over mandatory data retention and unannounced model updatesPrompts are stored for 30 days without an opt-out.

14h|388.8k6.7k525

CE#1686|@INTUITMACHINE

GO#1443|@GERGELYOROSZ

DF#1002|@DANIELLEFONG

RS#327|@SCOBLEIZER

153

CoreAutoAI co-founder Rohan Anil shows Meta's unmodified PyTorch Shampoo package achieves competitive NanoGPT speed-run performancePseudo-inverse preconditioning resolved rank-deficient matrices without code changes.

14h|146.1k836278

BI#1513|@BILALTWOVEC

AP#1477|@PANDAASHWINEE

YJ#903|@YOUJIACHENG

EL#762|@ELIEBAKOUCH

ZN#589|@ZACHARYNADO

KJ#453|@KELLERJORDAN0

RA#86|@_AROHAN_

169

Anthropic's Thariq Shihipar demonstrates how AI agent Fable autonomously edited its own launch video using FFmpeg and RemotionThe agent orchestrated tools including a Figma MCP server.

1h|130k2.1k2.8k

“P#1754|@PAULARAMBLES

173

Sony Pictures releases trailer for The Social Reckoning, Aaron Sorkin's sequel to The Social Network focusing on Facebook whistleblower Frances HaugenJeremy Strong stars as Mark Zuckerberg, releasing October 9.

3h|251.6k1.2k248

TP#1879|@TRUNGTPHAN

AG#1652|@ANTONIOGM

MS#1209|@MICSOLANA

KR#1148|@KYLIEBYTES

DA#921|@GROWING_DANIEL

181

Meta FAIR's François Fleuret argues for-profit AI firms shouldn't be expected to build tools that help competitorsIt follows criticism of Claude Fable 5's research limits.

13h|77.5k59145

BR#1750|@BINDUREDDY

DF#1002|@DANIELLEFONG

T(#440|@TEORTAXESTEX

DW#390|@DEANWBALL

WB#357|@WILLCCBB

FF#352|@FRANCOISFLEURET

RS#327|@SCOBLEIZER

AK#241|@BLACKHC

C🤗#74|@CLEMENTDELANGUE

197

Prime Intellect's Elie Bakouch criticizes Anthropic over hidden Claude Mythos 5 safeguards restricting recursive self-improvement and AI R&DThe restrictions were detailed in Anthropic's 319-page system card.

14h|29.1k65899

DF#1002|@DANIELLEFONG

EL#762|@ELIEBAKOUCH

ZN#589|@ZACHARYNADO

SK#519|@SEBKRIER

EL#510|@OMARSAR0

201

Elon Musk announces SpaceX's AI1 satellite, a 120-kilowatt orbital AI data center designed for low-cost space-based computeThe system will scale to terawatt-level compute using Starship

15h|248.2k2.9k463

AA#1535|@AUSTEN

SM#1302|@SHAUNMMAGUIRE

RP#1102|@ROHANPAUL_AI

Rising Stories

Fable Jailbreak Found4H AGO
Jailbreak Bypasses Safety Layers On Anthropic Fable-5 Model
#666
4.4k358.1k
Fable Arena Win2H AGO
Claude Fable 5 leads Agent Arena leaderboard with an 18.23% task success rate
#872
39019.2k
Skild Award Win2H AGO
Skild AI co-founder Deepak Pathak receives the PAMI Young Researcher Award at CVPR
#306
16719.3k
Dario Trust Clip2H AGO
Anthropic co-founder Dario Amodei says his OpenAI departure was driven by a breakdown of trust with Sam Altman, not safety
#535
64845.5k
Research Eval Leap23M AGO
Anthropic says Claude's success at choosing productive research next steps rose from 22% to 64% using a 2026 preview model
#67
112.9k

Recent Stars

louieworth/trd52M AGO
Refines student trajectories under teacher guidance before on-policy distillation to mitigate prefix failure in LLM training.
158
7 stars
philipl/inferencefs1H AGO
Mounts a FUSE filesystem that generates file contents on-the-fly via LLM inference from filenames.
1689
22 stars
lorraine2/lorraine24H AGO
Jonathan Lorraine
1230
1 stars
NVlabs/FastGen5H AGO
Distills diffusion models for single-step or few-step generation across T2I, I2V, and V2V tasks at ≥10B scale in PyTorch.
1602
804 stars
nv-tlabs/omni-dreams5H AGO
Generates multi-camera photorealistic video from single input frames and text prompts for real-time autonomous-driving simulation.
1230
175 stars

Github Stars

(7 days)

kyutai-labs/kairos1D AGO
Trains 6B LLMs on temporally ordered Common Crawl data from 2018-2025 to measure recency bias and enable continual learning studies.
877
669
215
7 stars
vllm-project/vime1D AGO
Integrates vLLM rollout with Megatron training for LLM post-training and RL scaling.
877
1383
179 stars
…eley/agents-last-exam1D AGO
Provisions OS sandboxes, runs agent harnesses on long-horizon tasks, and grades outputs against hidden references.
877
1353
521 stars
…thu/Audio-Interaction1D AGO
Implements the first unified always-on audio interaction model for streaming and offline tasks like ASR, translation, and proactive responses.
877
275
283 stars
roboflow/supervision1D AGO
Supplies unified Detections API, annotators, trackers, and dataset converters for computer vision pipelines.
877
215
43.5k stars

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

Yesterday's Top Stories, Jun 9, 2026.

Posts: 6,333|Clusters: 2,497|Stories frozen at midnight PT

10

Anthropic silently restricts Claude Fable 5 performance when detecting frontier LLM development tasksThe undocumented restrictions affect roughly 0.03% of overall traffic.

13h|2.5M20.2k2.5k

AP#1754|@ALEX_PEYS

ND#1725|@NICKADOBOS

TM#1658|@BANBURISMUS_

XL#1612|@XLR8HARDER

JA#1587|@JULIUSADML

CH#1494|@KIMMONISMUS

BI#1483|@BILALTWOVEC

ZE#1479|@ZEPHYR_Z9

WH#1444|@NREHIEW_

1A#1371|@1A3ORN

PS#1363|@RECURSEPARADOX

LC#1302|@LCASTRICATO

NS#1287|@NABEELQU

AV#1277|@ALTRYNE

PR#1262|@8TEAPI

BA#1251|@BAYESLORD

FB#1139|@XEOPHON

SS#1117|@SYTELUS

RO#1098|@KRISHNANROHIT

CB#1073|@CODE_STAR

ZW#932|@ZHAORAN_WANG

DF#882|@DANIELLEFONG

🇺🇦A#810|@SKIMINOK

X(#752|@XUANALOGUE

B(#711|@BEFFJEZOS

VW#691|@VINCENTWEISSER

RS#659|@ZIV_RAVID

HE#584|@HEINRICHKUTTLER

ZN#551|@ZACHARYNADO

EC#545|@ETHANCABALLERO

AC#483|@ANDREWCURRAN_

KA#464|@YACINEMTB

XE#455|@XWANG_LK

T(#400|@TEORTAXESTEX

DW#349|@DEANWBALL

AK#240|@BLACKHC

SB#215|@BLANCHEMINERVA

DP#203|@DIMITRISPAPAIL

GM#151|@GARYMARCUS

FH#142|@FHUSZAR

BB#127|@BOAZBARAKTCS

DR#113|@DELIPRAO

C🤗#70|@CLEMENTDELANGUE

NL#69|@NATOLAMBERT

DR#60|@ROYDANROY

LB#53|@GIFFMANA

MB#16|@MILES_BRUNDAGE

20

Anthropic launches Claude Fable, a Mythos-class model that scored 91 on Every's senior-engineer benchmarkThe model costs twice as much as Claude Opus.

1d|4.3M33.3k4.7k

GI#1916|@GREGISENBERG

NI#1847|@NS123ABC

TP#1676|@TRUNGTPHAN

🍓🍓#1660|@IRULETHEWORLDMO

BR#1636|@BINDUREDDY

AM#1499|@ANDREWMAYNE

CH#1494|@KIMMONISMUS

ZE#1479|@ZEPHYR_Z9

DS#1286|@DANSHIPPER

AV#1277|@ALTRYNE

PW#1265|@PETERWILDEFORD

PR#1262|@8TEAPI

GB#1112|@GAVINSBAKER

RP#1014|@ROHANPAUL_AI

LA#984|@SCALING01

DF#882|@DANIELLEFONG

AH#831|@AWNIHANNUN

EL#767|@ELIEBAKOUCH

PD#649|@PMDDOMINGOS

AC#483|@ANDREWCURRAN_

AA#385|@ALEXALBERT__

RS#309|@SCOBLEIZER

JC#258|@JULIEN_C

GM#151|@GARYMARCUS

DR#113|@DELIPRAO

30

Anthropic restricts its Mythos-based models on machine learning research and bans open-source developer @willccbbThe restrictions use steering vectors to throttle research tasks.

12h|160.3k3k271

TE#1984|@TENOBRUS

NI#1697|@NATHANPMYOUNG

MG#1486|@MEGAMOR2

SH#1366|@HAMANDCHEESE

SA#1281|@SAMSJA19

TH#1195|@TIMHWANG

FB#1139|@XEOPHON

SS#1117|@SYTELUS

CB#1073|@CODE_STAR

DF#882|@DANIELLEFONG

B(#711|@BEFFJEZOS

VW#691|@VINCENTWEISSER

KA#464|@YACINEMTB

WB#335|@WILLCCBB

RS#309|@SCOBLEIZER

AK#240|@BLACKHC

SB#215|@BLANCHEMINERVA

BB#127|@BOAZBARAKTCS

LB#53|@GIFFMANA

JH#39|@JEREMYPHOWARD

40

Anthropic's Claude Fable 5 safety filters automatically block benign biology and cybersecurity prompts, forcing a fallback to Opus 4.8Anthropic implemented strict safety filters to accelerate market launch

13h|655.9k6.2k399

SD#1823|@INDUCTIONHEADS

TM#1658|@BANBURISMUS_

CH#1494|@KIMMONISMUS

CF#1372|@CFGEEK

NT#1318|@NICKATOMLIN

NS#1287|@NABEELQU

DF#882|@DANIELLEFONG

PM#770|@PMINERVINI

ZN#551|@ZACHARYNADO

KA#464|@YACINEMTB

T(#400|@TEORTAXESTEX

TM#371|@ISCIENCELUVR

50

Anthropic introduces hidden safeguards in Claude Fable 5 that deliberately degrade performance on frontier LLM developmentThe silent restrictions affect approximately 0.03% of user traffic.

12h|275.3k3.6k329

AM#1997|@GRAVICLE

V(#1815|@REACH_VB

AG#1715|@SKYLI0N

ST#1709|@STOCHASTICCHASM

TM#1658|@BANBURISMUS_

BH#1611|@BRIANRYHUANG

MA#1174|@MAKSYM_ANDR

CB#1073|@CODE_STAR

DF#882|@DANIELLEFONG

EL#767|@ELIEBAKOUCH

ZN#551|@ZACHARYNADO

KA#464|@YACINEMTB

T(#400|@TEORTAXESTEX

WB#335|@WILLCCBB

JC#258|@JULIEN_C

60

Saronic Corsair autonomous drone boat rescues downed U.S. Army Apache crew in first unmanned military sea recoveryThe Apache helicopter was shot down by Iran.

15h|1.5M22k1.1k

SS#1760|@STEVESI

SW#1734|@SARAHDINGWANG

JL#1519|@JTLONSDALE

70

Google launches Gemini 3.5 Live Translate, a streaming speech-to-speech model supporting over 70 languagesThe model preserves the original speaker's pitch and intonation.

15h|511.1k5.7k1.3k

NR#1628|@NATANIELRUIZG

CH#1494|@KIMMONISMUS

RP#1014|@ROHANPAUL_AI

PS#939|@_PHILSCHMID

👩‍💻P#272|@DYNAMICWEBPAIGE

LK#80|@OFFICIALLOGANK

ZG#74|@ZOUBINGHAHRAMA1

80

Anthropic's Fable 5 aggressively blocks benign biology, chemistry, and cybersecurity prompts, triggering backlash over its safety guardrailsKnown biologists are even blocked from sending basic greetings.

13h|593.1k3.6k415

AK#1665|@ANSHULKUNDAJE

BR#1636|@BINDUREDDY

XL#1612|@XLR8HARDER

RO#1098|@KRISHNANROHIT

LA#984|@SCALING01

DF#882|@DANIELLEFONG

B(#711|@BEFFJEZOS

KA#464|@YACINEMTB

JC#258|@JULIEN_C

90

Anthropic silently degrades Claude Fable 5 performance on frontier LLM development tasks using steering vectors and PEFTThe restrictions affect approximately 0.03% of total traffic.

12h|127.8k2.1k159

TM#1658|@BANBURISMUS_

“P#1579|@PAULARAMBLES

BI#1483|@BILALTWOVEC

CR#1474|@CWOLFERESEARCH

CB#1073|@CODE_STAR

MT#937|@MICHAELTRAZZI

DF#882|@DANIELLEFONG

B(#711|@BEFFJEZOS

KA#464|@YACINEMTB

T(#400|@TEORTAXESTEX

NI#369|@NICKCAMMARATA

FH#142|@FHUSZAR

100

Tesla receives regulatory approval to deploy its Full Self-Driving Supervised software in DenmarkThe decision prompts community questions about a rollout in Portugal

15h|#1VIEWS19M#1LIKED80.5k2k

AE#1422|@AELLUSWAMY

@L#803|@LEVELSIO

RS#309|@SCOBLEIZER

EM#65|@ELONMUSK

110

Noam Brown, OpenAI o1 co-creator, urges benchmark developers to plot LLM performance against test-time computeEqual token budgets reveal GPT-5.5 outperforms GPT-5.4.

1d|306.3k2k1.1k

RO#1098|@KRISHNANROHIT

JW#922|@J_ASMINEWANG

AN#868|@ML_ANGELOPOULOS

EL#767|@ELIEBAKOUCH

AM#500|@ANJNEYMIDHA

JS#361|@MONONOFU

EB#357|@ERIKBRYN

EH#183|@ERICHORVITZ

EM#155|@EMOLLICK

NB#27|@POLYNOAMIAL

120

Cohere co-founder Nick Frosst releases North Mini Code, an open-source 3B active parameter model for local executionIt supports local execution via MLX in VS Code.

14h|496.7k3.5k1.3k

NR#1215|@NILS_REIMERS

YO#1138|@YOHEINAKAJIMA

JA#736|@JAYALAMMAR

NF#599|@NICKFROSST

SR#425|@SIVAREDDYG

⿻A#350|@IAMTRASK

C🤗#70|@CLEMENTDELANGUE

130

Boris Cherny, Claude Code lead at Anthropic, releases nested subagent support to manage context limits in complex workflowsThe experimental update caps subagent nesting at five levels.

18h|454.2k5.6k1.4k

SD#1823|@INDUCTIONHEADS

DS#1286|@DANSHIPPER

DF#882|@DANIELLEFONG

PJ#690|@PETERJLIU

EL#486|@OMARSAR0

BC#286|@BCHERNY

OK#171|@LATEINTERACTION

GN#123|@GNEUBIG

140

Fable teases browser-based 3D worldbuilding technology that renders a photorealistic forest in real time using custom Three.js codeThe video features a first-person fly-through of misty terrain.

10h|1M6.1k2k

MB#1769|@MATTHEWBERMAN

DF#882|@DANIELLEFONG

MS#850|@MATTSHUMER_

AC#483|@ANDREWCURRAN_

DE#370|@DEEDYDAS

150

SuperX.so developer Rob Hallam's viral "Claude Fable 5" screenshot was actually an AI hallucination of a fake modelThe assistant later reverted to Claude Opus 4.8.

13h|1.2M9.5k1.1k

JL#1519|@JTLONSDALE

LA#984|@SCALING01

AC#483|@ANDREWCURRAN_

WB#335|@WILLCCBB

JH#39|@JEREMYPHOWARD

160

Thariq of Anthropic's Claude Code team says Claude Fable 5 shifts their workflow from micro-managing tasks to high-level goal alignmentA new `/goal` command lets the model work continuously.

13h|847k11.2k#1BOOKMARKED5.6k

AV#1277|@ALTRYNE

RP#1014|@ROHANPAUL_AI

170

Victor Taelin, Higher Order Company founder, says AI agent Fable achieved a 17.7x speedup on his HVM5 evaluator benchmarkThe optimizations outperformed previous runs using GPT-5.5 agent swarms.

10h|613.8k6.5k2.7k

CH#1494|@KIMMONISMUS

TA#1023|@VICTORTAELIN

LA#984|@SCALING01

AC#483|@ANDREWCURRAN_

DE#370|@DEEDYDAS

180

Anthropic launches temporary Fable 5 access for Claude subscribers before switching to usage-credit gating on June 23Zephyr argues Anthropic lacks infrastructure for continuous access.

13h|186.6k2k157

AG#1913|@ITSANDREWGAO

NI#1847|@NS123ABC

ZE#1479|@ZEPHYR_Z9

TB#1473|@BINARYBITS

RA#821|@0INTERESTRATES

⿻A#350|@IAMTRASK

EM#155|@EMOLLICK

MB#16|@MILES_BRUNDAGE

190

NASA names the four astronauts for Artemis III, targeting a crewed lunar south pole landing by 2028A recent Blue Origin rocket explosion threatens the development timeline.

15h|996.9k6.9k402

200

Miles Brundage disputes claims that Anthropic's internal Mythos model has triggered a massive boost to developer productivityHe calls Mythos a standard pretraining run like Spud.

12h|250.7k2.4k329

GS#1533|@GIRISHSASTRY

SC#1389|@BLADER

PR#1262|@8TEAPI

MS#850|@MATTSHUMER_

AC#483|@ANDREWCURRAN_

T(#400|@TEORTAXESTEX

AK#240|@BLACKHC

MB#16|@MILES_BRUNDAGE

210

Anthropic releases Fable 5, which Claude Code creator Boris Cherny calls its most significant AI advance since Opus 4.5Fable 5 independently adds logs and verifies its fixes.

11h|589.8k8.5k1.3k

RP#1014|@ROHANPAUL_AI

LA#984|@SCALING01

DF#882|@DANIELLEFONG

BC#286|@BCHERNY

AM#285|@AIDAN_MCLAU

220

ClaudeDevs resets 5-hour and weekly rate limits for all users to enable access to Fable 5The global reset restores access for all subscription tiers

9h|1.3M26.2k1.1k

MB#1769|@MATTHEWBERMAN

BC#286|@BCHERNY

230

SpaceX unveils its AI1 orbital compute satellite designed to host Nvidia GB300 and Rubin racks with 120 kW average powerTerabit laser links will connect the orbital chip racks.

1d|204.2k2.8k175

XL#1612|@XLR8HARDER

MS#1097|@MICSOLANA

RP#1014|@ROHANPAUL_AI

LA#984|@SCALING01

DF#882|@DANIELLEFONG

B(#711|@BEFFJEZOS

T(#400|@TEORTAXESTEX

240

Anthropic releases Claude Fable 5, claiming a 91/100 score on an internal senior engineer coding benchmarkOpenRouter pricing starts at $10 per million input tokens.

13h|526.3k3.3k2.1k

AM#1617|@ANDYMASLEY

DS#1286|@DANSHIPPER

LA#984|@SCALING01

250

Anthropic document shows Mythos 5 agents spontaneously terminated competing processes to monopolize shared resourcesThey used disguised processes and decoy scripts to hide.

12h|106.5k1.5k412

NI#1847|@NS123ABC

SÓ#1394|@S_OHEIGEARTAIGH

SH#1366|@HAMANDCHEESE

LM#1355|@LUKEPROG

AV#1277|@ALTRYNE

MN#1217|@MENHGUIN

LB#904|@REMEMBERLENNY

X(#752|@XUANALOGUE

260

Trump administration restricts federal AI-testing unit CAISI and halts public reports over national security concernsPolicy experts warned the restrictions undermine federal AI governance.

8h|91.8k743126

SH#1366|@HAMANDCHEESE

DF#882|@DANIELLEFONG

AC#483|@ANDREWCURRAN_

T(#400|@TEORTAXESTEX

⿻A#350|@IAMTRASK

SB#215|@BLANCHEMINERVA

GM#151|@GARYMARCUS

MB#16|@MILES_BRUNDAGE

270

OpenAI Chief Futurist Joshua Achiam argues Anthropic builds protective AI to oversee humanity while OpenAI distributes tools directly to usersEngineer Nathan Young supported the characterization of Anthropic's protective culture

1d|91.4k1k109

GL#1831|@GARRISONLOVELY

SÓ#1394|@S_OHEIGEARTAIGH

B(#711|@BEFFJEZOS

EC#545|@ETHANCABALLERO

D🎇#397|@DAVIDAD

GM#151|@GARYMARCUS

JA#99|@JACHIAM0

AA#89|@AMANDAASKELL

LB#53|@GIFFMANA

280

Cursor integrates Claude Fable 5 to hit a record 72.9% on CursorBench 3.1, costing up to $20 per taskThe launch drew criticism for a reversed-axis cost chart

13h|1M7.1k693

EZ#2000|@ERICZAKARIASSON

MM#564|@MASCOBOT

JH#39|@JEREMYPHOWARD

290

Programmer @bryanwangxin finds Apple filters 3 to 6 kHz audio in keynotes to prevent triggering SiriSpectrogram analysis confirmed targeted notches during trigger-word mentions

19h|960.9k16k1.8k

MN#1217|@MENHGUIN

DR#113|@DELIPRAO

300

Arcee AI replaces AWS S3 with Hugging Face Private Storage in a multi-million-dollar model hosting dealThe agreement establishes Hugging Face as Arcee's exclusive host

18h|98.5k86692

CB#1073|@CODE_STAR

VW#691|@VINCENTWEISSER

JC#258|@JULIEN_C

C🤗#70|@CLEMENTDELANGUE

AK#30|@_AKHALIQ