Tech News

Today's Highlights

In case you missed it

US orders Anthropic to suspend foreign access to Fable 5 and Mythos 5 over jailbreak concerns

NL#80|@NATOLAMBERT

MB#57|@MILES_BRUNDAGE

#1 VIEWED

Y Combinator co-founder Paul Graham publishes 'How to Earn a Billion Dollars' essay detailing startup wealth creation

KB#1038|@KTMBOYLE

MP#1028|@EASTDAKOTA

AO#636|@ALEXISOHANIAN

FASTEST CLIMBING

Meta's Lucas Beyer notes AI startup logos are shifting from circular emblems to wave-like and chevron patterns

CB#1674|@CODE_STAR

EM#171|@EMOSTAQUE

LB#72|@GIFFMANA

RISING LIKES

233

China's GLM 5.2 launches and takes the top BridgeBench spot, running at 300 tokens per second at one-tenth the cost of rivals

BR#1879|@BINDUREDDY

RS#447|@SCOBLEIZER

Top Stories

Posts:—|Clusters:—|Next Crawl:—

11

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

4h|#1VIEWS459.5k#1LIKED4.1k#1BOOKMARKED4.7k#1COMMENTS328

KB#1038|@KTMBOYLE

MP#1028|@EASTDAKOTA

AO#636|@ALEXISOHANIAN

Top comment: @paulg “@aaravshirpurkar Interestingly, I don't think Claude could do it, because the instructions are to start by making something you and your friends want, and Claude neither has friends nor (as far as I know) wants anything.”

21

Ross Taylor, former Meta reasoning lead, argues DeepMind's dominance historically hindered independent UK AI startupsEmad Mostaque argued that Nvidia's Nemotron stack simplifies model pretraining

7h|78.6k553289

MC#1529|@MATTHEWCLIFFORD

RO#1315|@KRISHNANROHIT

RT#842|@ROSSTAYLOR90

NB#353|@NATHANBENAICH

EM#171|@EMOSTAQUE

Top comment: @peterrhague “@rosstaylor90 The UK also has the advantage of being an attractive place to work for all the people just locked out of their own models for their citizenship status. For all the moaning, London ranks quite highly among mobile cosmopolitan types.”

30

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

6h|50k626145

V(#1939|@REACH_VB

MN#1716|@MENHGUIN

GA#598|@GABRIEL1

B(#530|@BEFFJEZOS

Top comment: @TheAhmadOsman “@beffjezos Why not both? :)”

Watchlist

Video Signals

RP#1257|@ROHANPAUL_AI

Vinod Khosla Prefers Full AI Replacements Over Co-Pilots to Bypass Human Resistance

7h ago|Views 21KLikes 102Bookmarks 51

EM#184|@EMOLLICK

AI Generates Interactive FTL Travel Simulator From One Prompt

2h ago|Views 21KLikes 320Bookmarks 127

RP#1257|@ROHANPAUL_AI

Sam Altman Says Programming Loses Value as AI Era Favors Adaptability

7h ago|Views 455KLikes 114Bookmarks 60

BA#713|@ADCOCK_BRETT

Humanoid Robot Descends Stairs in Steady Forward Gait

49m ago|Views 7KLikes 262Bookmarks 14

EM#52|@ELONMUSK

Tesla FSD Reacts To Thrown Dummy In 130 Milliseconds

7h ago|Views 100KLikes 0Bookmarks 0

RP#1257|@ROHANPAUL_AI

Meituan Deploys Autonomous Delivery Robots on Chinese City Streets

6h ago|Views 93KLikes 97Bookmarks 21

RP#1257|@ROHANPAUL_AI

Vinod Khosla Warns AI Will Eliminate India's Traditional BPO And IT Services

5h ago|Views 21KLikes 30Bookmarks 12

KD#904|@CSPROFKGD

AI Researcher Stays Passive With Each Frontier Model Release

6h ago|Views 380KLikes 30Bookmarks 5

RP#1257|@ROHANPAUL_AI

Dreamina Prepares Seedance 2.0 Mini Release for June 15

1h ago|Views 1.4KLikes 2Bookmarks 1

XL#1855|@XLR8HARDER

Builder Demonstrates Real-Time Visualization For AI Agent Workflows

8h ago|Views 532Likes 3Bookmarks 1

40

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

1h|249.1k2k184

AV#1419|@ALTRYNE

SI#584|@SIGNULLL

Top comment: @salvomic “@markgurman Aboslutely. However, there is an important thing to say: if Apple does not resolve the issues with the EU and does not release Siri AI for iOS in Europe soon, it will lose many users.”

Digg Deeper question: “How much of this is Apple models vs Google?”

51

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

3h|93.4k2.2k247

RH#1332|@RRHOOVER

Top comment: @ryemike_merio “@paulg If you don't have a degree and ten years of experience, it could just be an honest statement. There are non-technical founders who are much smarter than PHD's.”

New github stars (48hrs)

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

omnigent-ai/omnigent9711d

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

62

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

3h|34.4k61252

T(#501|@TEORTAXESTEX

KA#403|@YACINEMTB

Top comment: @teortaxesTex “little warmth here”

73

Analyst Zephyr argues new export controls are constrained by Nvidia and hyperscaler reliance on Chinese AI hardware componentsShortages forced Nvidia to qualify Chinese PCB and optical suppliers

4h|65.1k612285

ZE#1695|@ZEPHYR_Z9

T(#501|@TEORTAXESTEX

Top comment: @teortaxesTex “@zephyr_z9 Impressive how the operational logic is "we can attack with impunity. There are no costs beyond losing some market share. Why aren't we attacking all the time then? I notice I am confused… perhaps it's treason". Not a thought about actual limits on available options.”

83

Meta's Lucas Beyer notes AI startup logos are shifting from circular emblems to wave-like and chevron patternsFounder Emad Mostaque responded humorously to the design observation.

4h|11.5k1156

FASTEST CLIMBING+3

CB#1674|@CODE_STAR

EM#171|@EMOSTAQUE

LB#72|@GIFFMANA

Top comment: @A_K_Nain “@giffmana Trying to build AGI, and can't design a unique, subtle, and gorgeous logo is the irony!”

90

Hardware analyst Zephyr says AI workloads and HBM demand have reversed a 63-year trend of declining DRAM costsDRAM costs previously declined tenfold every five years.

3h|38.4k22165

MN#1716|@MENHGUIN

ZE#1695|@ZEPHYR_Z9

Top comment: @lloydz3 “@jukan05 💵Current stock allocations are listed below ⬇️🔻 👇🔽”

102

Study finds LLM agents ignore condensed self-improvement rules, relying entirely on raw task historiesCorrupting raw history logs collapses agent performance.

1h|12.5k14170

RP#1257|@ROHANPAUL_AI

GM#178|@GARYMARCUS

Top comment: @JSBach_007 “@NewsCop9000 @GaryMarcus @dwarkesh_sp You sound right to me. I think AI is ultimately just an idol onto which people project their hopes and dreams. There is an actual technology there, but it isn't capable of 1/10th of what they project. The fundamental nature of the tech is misunderstood, fixed and overstated.”

112

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

1h|16.3k10264

RP#1257|@ROHANPAUL_AI

GT#121|@GARRYTAN

Top comment: @rickernel “@rohanpaul_ai Interesting paper. Send this to Amodei.”

127

China's GLM 5.2 launches and takes the top BridgeBench spot, running at 300 tokens per second at one-tenth the cost of rivalsIt debuted after the US banned Claude Fable 5.

6h|9.4k2339

RISING LIKES233

BR#1879|@BINDUREDDY

RS#447|@SCOBLEIZER

Top comment: @bridgebench “@bridgemindai GLM 5.2 is the best Chinese open source model that we have tested.”

131

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

8h|32.3k9811

LA#1215|@SCALING01

T(#501|@TEORTAXESTEX

Top comment: @boneGPT “@scaling01 strong circle”

147

Yann LeCun argues Anthropic's current challenges are a direct consequence of its prior actionsAcademics Yi Ma and Kosta Derpanis endorsed his critique.

6h|4.4k491

KD#904|@CSPROFKGD

YM#423|@YIMATWEETS

Top comment: @midnightbobarun “@mark_k @AnthropicAI It was very polite of Yann LeCun to not just write "FAFO" about Anthropic 🤭”

150

Bojan Tunguz, TabulAI founder, argues that AI issues will eventually eclipse all other political topicsFounder Tanishq Mathew Abraham opposed the projected transition.

5h|3.8k521

TM#613|@ISCIENCELUVR

Top comment: @danfaggella “@tunguz You speak of the political singularity, friend https://danfaggella.com/political-singularity”

161

Competitor "gum" sets 1,097-microsecond record in GPU MODE B200 challenge amid exploit debateThe contest optimizes batched QR decomposition kernels.

7h|3.6k314

RA#102|@_AROHAN_

LB#72|@GIFFMANA

Top comment: @_arohan_ “@giffmana previous 3 did some clever exploit like overwrite timing is my guess”

171

Microsoft Research's Dimitris Papailiopoulos uses Opus 4.6 and Codex 5.5 xhigh as fallbacks when Fable is offlineHe considers Opus 4.6 the top-performing 4.x model.

2h|3k258

FB#1778|@XEOPHON

DP#217|@DIMITRISPAPAIL

Top comment: @xeophon “@DimitrisPapail well well well”

180

Arcee AI's Cody Blakeney argues @kalomaze's new benchmark should test reasoning budgets instead of disabling reasoningDeveloper @kalomaze designed the evaluation to measure ungameable scaling advantages.

6h|88551

CB#1674|@CODE_STAR

AM#359|@AIDAN_MCLAU

190

OpenRouter's $1.5 million daily Anthropic revenue highlights the scale gap with direct enterprise contractsTotal OpenRouter token volume continues to grow significantly.

3h|388141

SO#1919|@CIS_FEMALE

CB#1674|@CODE_STAR

Top comment: @cis_female “@code_star @giffmana @WangUWS @OpenRouter OpenRouter reports ~$1.5M/day in Anthropic revenue vs Anthropic iirc $40B/y ARR they claimed = $109M/day”

200

Arcee AI's Cody Blakeney questions whether EU regulatory delays on AI are due to summer vacationFlorian Brand agreed that European policymakers are sluggish.

2h|1.1k160

FB#1778|@XEOPHON

CB#1674|@CODE_STAR

Top comment: @xeophon “@code_star Yes”

Rising Stories

China Axes Degrees5H AGO
China eliminates 12,000 university degree programs to restructure its higher education system for the AI era
#423
6.5k475.6k
Jailbreak Layers Revealed7H AGO
Susan Zhang maps how LLM reasoning layers also enable jailbreaks
#84
57568.4k
Compound Models Win4H AGO
Trask and Manheim debate why ensembling beats single frontier models
#595
43663k

Recent Stars

Emanuele-web04/synara3H AGO
Unifies multiple AI coding subscriptions into one local-first desktop workspace with chats, terminals, and previews.
1404
825 stars
shadcn/improve3H AGO
Audits codebases with a capable model and generates prioritized implementation plans for cheaper agents to execute.
1404
3.9k stars
…r-turner/claude-guard11H AGO
Wraps Claude Code sessions with deterministic checks on commands and writes to limit agent access.
570
11 stars
alanagoyal/alanagoyal20H AGO
Implements a macOS-inspired personal website with interactive desktop environment and apps.
278
301 stars
…gacel/auto-gpu-kernel20H AGO
Autonomously discovers and optimizes GPU kernels for sparse attention workloads.
754
115 stars

Github Stars

(7 days)

macrodata-labs/refiner2D AGO
Processes and refines large-scale ML datasets via a data framework.
1136
353
888
1674
41 stars
omnigent-ai/omnigent1D AGO
Implements a domain-agnostic framework for autonomous AI agents with ReAct loops, multi-LLM routing, reasoning graphs and MCP integration.
586
1066
971 stars
vllm-project/vime4D AGO
Integrates vLLM rollout with Megatron training for LLM post-training and RL scaling.
1014
1768
213 stars
…eley/agents-last-exam4D AGO
Provisions OS sandboxes, runs agent harnesses on long-horizon tasks, and grades outputs against hidden references.
1014
1489
521 stars
datacurve-ai/deep-swe4D AGO
Benchmarks frontier coding agents on 113 long-horizon tasks from active open-source repos with isolated environments and program verifiers.
1014
1853
761 stars

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

Yesterday's Top Stories, Jun 13, 2026.

Posts: 3,333|Clusters: 1,208|Stories frozen at midnight PT

10

US orders Anthropic to suspend foreign access to Fable 5 and Mythos 5 over jailbreak concernsOther Claude models remain unaffected by the export directive

1d|9.4M48.9k#1BOOKMARKED15.3k

PS#1594|@RECURSEPARADOX

MB#1589|@BAKKERMICHIEL

NX#1568|@NXTHOMPSON

AD#1537|@DORIALEXANDER

SS#1471|@SYTELUS

AN#1431|@ARAN_NAYEBI

RR#1394|@RAJKO_RAD

CH#1360|@KIMMONISMUS

RO#1315|@KRISHNANROHIT

RP#1257|@ROHANPAUL_AI

C🎉#1248|@CHARLES_IRL

AP#1222|@APOMPLIANO

LA#1215|@SCALING01

GO#1155|@GERGELYOROSZ

JL#1086|@JESSICALESSIN

SF#1000|@FEIZISOHEIL

ZN#983|@ZACHARYNADO

AK#982|@ANDREY_KURENKOV

TC#966|@TUHINCHAKR

AG#924|@ARGLEAVE

SS#819|@STEVESI

DF#756|@DANIELLEFONG

RS#741|@ZIV_RAVID

AC#682|@ANDREWCURRAN_

SK#667|@SEBKRIER

GA#598|@GABRIEL1

⿻A#595|@IAMTRASK

MS#585|@MARTINSHKRELI

B(#530|@BEFFJEZOS

T(#501|@TEORTAXESTEX

RS#447|@SCOBLEIZER

KA#403|@YACINEMTB

AV#350|@ASHLEEVANCE

CP#347|@CHAMATH

T🪽#331|@TEKNIUM

MA#315|@MARTIN_CASADO

AB#228|@ABEIRAMI

GK#197|@THEGAUTAMKAMATH

GM#178|@GARYMARCUS

GT#121|@GARRYTAN

DR#113|@DELIPRAO

NL#80|@NATOLAMBERT

MB#57|@MILES_BRUNDAGE

20

US export controls force Anthropic to disable Fable 5 and Mythos 5 models globally after a successful jailbreakThe restricted models contain sensitive CBRN and offensive cyber data

1d|509.8k5k705

MB#1882|@MATTHEWJBAR

BR#1861|@BRIANROEMMELE

ZE#1695|@ZEPHYR_Z9

JA#1672|@JULIUSADML

RR#1394|@RAJKO_RAD

CH#1360|@KIMMONISMUS

RO#1315|@KRISHNANROHIT

RP#1257|@ROHANPAUL_AI

LA#1215|@SCALING01

SA#1139|@SAMUELALBANIE

MP#1028|@EASTDAKOTA

RT#842|@ROSSTAYLOR90

JT#796|@TOGELIUS

DF#756|@DANIELLEFONG

PM#714|@PMINERVINI

YJ#586|@YUCHENJ_UW

SI#584|@SIGNULLL

CT#542|@CHENHAOTAN

RS#447|@SCOBLEIZER

AK#271|@BLACKHC

GM#178|@GARYMARCUS

@(#155|@TIMNITGEBRU

Digg Deeper question: “Impact and consequences of this”

30

Trump administration gave Anthropic 90 minutes to pull Claude Mythos 5 and Fable 5 after Amazon warned of security risksThe action prompted discussions of new U.S. export controls

14h|2M7.7k1.6k

XL#1855|@XLR8HARDER

MB#1780|@MATTHEWBERMAN

FB#1778|@XEOPHON

ZE#1695|@ZEPHYR_Z9

CH#1360|@KIMMONISMUS

JL#1086|@JESSICALESSIN

RK#1011|@MIKEISAAC

DS#975|@DANSHIPPER

TC#966|@TUHINCHAKR

AC#682|@ANDREWCURRAN_

T🪽#331|@TEKNIUM

GM#178|@GARYMARCUS

Digg Deeper question: “What was the security flaw?”

40

Anthropic founder Dario Amodei proposes a regulatory framework requiring third-party safety testing and government power to block unsafe frontier AI modelsThe framework targets bioweapons, cybersecurity, and automated R&D.

3d|171.9k2.3k208

BR#1879|@BINDUREDDY

ST#1694|@_ONIONESQUE

NS#1570|@NABEELQU

HB#1424|@HERBIEBRADLEY

RO#1315|@KRISHNANROHIT

YP#1171|@YPATIL125

AK#982|@ANDREY_KURENKOV

MK#869|@MIKEKNOOP

DF#756|@DANIELLEFONG

B(#530|@BEFFJEZOS

WM#520|@WILLMANIDIS

MA#315|@MARTIN_CASADO

MB#57|@MILES_BRUNDAGE

Digg Deeper question: “Summarize Dario's take”

50

SpaceX has launched more satellites into orbit than all other space programs in history combinedElon Musk projects a fivefold lead within five years.

1d|16.2M111.9k3.3k

MA#1240|@ARRINGTON

JC#943|@JASONCRAWFORD

EM#52|@ELONMUSK

60

OpenRouter launches Fusion API, parallel-routing queries to beat solo frontier models on Perplexity's DRACO benchmarkThe system matches Fable 5 deep research performance at half the cost

12h|199.2k2k548

TS#1677|@TIMSHI_AI

ND#1653|@NICKADOBOS

AA#1510|@ALEXATALLAH

VW#990|@VINCENTWEISSER

DF#756|@DANIELLEFONG

EL#684|@OMARSAR0

⿻A#595|@IAMTRASK

B(#530|@BEFFJEZOS

T(#501|@TEORTAXESTEX

AM#437|@ANJNEYMIDHA

KA#403|@YACINEMTB

EM#171|@EMOSTAQUE

Digg Deeper question: “so, the tldr is that the same request hits multiple models then a judge sorts through the mess, the judge uses a synthesizer, who wrote the synthesizer and what model is it using?”

70

Rio de Janeiro's municipal IT company releases Rio 3.5 Open 397B, an open-weights LLM that beats Qwen 3.7 Plus on Terminal 2.1The model uses dynamic SwiReasoning to optimize token efficiency.

9h|397.6k5.1k1.1k

CH#1360|@KIMMONISMUS

YP#1171|@YPATIL125

AC#682|@ANDREWCURRAN_

T(#501|@TEORTAXESTEX

RS#447|@SCOBLEIZER

DP#174|@DYLAN522P

C🤗#109|@CLEMENTDELANGUE

80

Elon Musk's comment on Starship's massive scale prompts e/acc founder Beff to recall the rocket's original 'BFR' codenameBFR refers to Big Falcon Rocket, Starship's original name.

22h|#1VIEWS17.7M#1LIKED153.6k3.4k

B(#530|@BEFFJEZOS

EM#52|@ELONMUSK

90

Anthropic disputes reports that Amazon-discovered jailbreaks in its Claude model triggered federal distribution restrictionsAnthropic dismissed the findings as minor, previously known bugs.

22h|1M7.2k1.4k

CH#1360|@KIMMONISMUS

T🪽#331|@TEKNIUM

100

Z.ai launches GLM-5.2 with a 1-million-token context window ahead of an MIT-licensed release next weekInitial access is limited to GLM Coding Plan subscribers.

22h|409.3k4.8k656

DC#1803|@VICTOR207755822

AJ#1798|@ALEXJC

ZE#1695|@ZEPHYR_Z9

CH#1360|@KIMMONISMUS

T(#501|@TEORTAXESTEX

KA#403|@YACINEMTB

110

Anthropic suspends Fable 5 after a safeguard rollback let the model triple Gemma 4 WebGPU inference speedThe suspension reset the model's benchmark scores to zero.

1d|344.9k3.5k800

MN#1716|@MENHGUIN

CH#1360|@KIMMONISMUS

GK#1219|@GREGKAMRADT

TA#1140|@VICTORTAELIN

AC#682|@ANDREWCURRAN_

KA#403|@YACINEMTB

120

Databricks founder Matei Zaharia open-sources Omnigent, a meta-harness to orchestrate and govern multiple AI agents without model lock-inIt supports Model Context Protocols and automated session failover

14h|193.5k1.6k1.2k

AG#1613|@ALIGHODSI

AD#1066|@MRDROZDOV

EL#684|@OMARSAR0

YJ#586|@YUCHENJ_UW

MZ#441|@MATEI_ZAHARIA

130

Europe 2031 report warns of economic decline as three US labs each hold more AI compute than the continentThe analysis projects a 12.3-fold US compute advantage by 2031

2d|77.6k813103

ZE#1695|@ZEPHYR_Z9

MB#1589|@BAKKERMICHIEL

SÓ#1563|@S_OHEIGEARTAIGH

AD#1537|@DORIALEXANDER

T(#501|@TEORTAXESTEX

KY#414|@KAYO_YIN

AK#271|@BLACKHC

140

TrainXGB creator Bojan Tunguz jokes that an organization is hiding a highly capable, "Mythos-level" AI modelThe joke featured an Inglourious Basterds interrogation meme.

17h|411.8k14k626

NB#239|@NIKITABIER

150

SpaceX's equity compensation model has reportedly made up to 5,000 employees millionaires, including blue-collar and non-technical staffNearly 400 staff members will exceed $100 million.

2d|305.1k5.7k328

MA#1240|@ARRINGTON

KB#1038|@KTMBOYLE

160

Google brings its legacy Earth Pro Flight Simulator to the web, bypassing high 3D tile API costsThe feature was previously restricted to the desktop-only Earth Pro.

21h|699.3k3.3k1.6k

BS#1756|@BILAWALSIDHU

DF#756|@DANIELLEFONG

@L#627|@LEVELSIO

170

AI Safety Fears Expand Lab Power Into Cyberpunk Chaebol Future

7h|118.7k1.5k647

XL#1855|@XLR8HARDER

180

Alex Rives and UC Berkeley researchers develop a laser phase plate for atomic-resolution cryo-electron tomographyIt resolves tiny proteins and molecular machines inside cells.

2d|167.8k996342

CK#950|@KALINOWSKI007

DF#756|@DANIELLEFONG

AV#350|@ASHLEEVANCE

BJ#316|@BRYAN_JOHNSON

190

Mark Saroufim launches a GPU MODE competition to optimize batched QR decomposition on NVIDIA B200 GPUsEarly submissions show a 260x speed gap between leaders.

1d|91.4k420144

EL#1136|@ELIEBAKOUCH

MS#1069|@MARKSAROUFIM

ZN#983|@ZACHARYNADO

KA#403|@YACINEMTB

RA#102|@_AROHAN_

200

Matan Grinberg, Factory co-founder and CEO, warns enterprises face an unsustainable "token hangover" from overusing expensive frontier modelsHe proposes routing 80% of workloads to open-source models.

14h|286k6261.4k

GR#1176|@GOKULR

HS#800|@HARRYSTEBBINGS

210

Forward Future's Matthew Berman and tech executive Garry Tan urge adopting local open-source AI to avoid proprietary lock-inTan highlighted tools like Hermes and OpenClaw for independence.

11h|105.2k1.1k771

MB#1780|@MATTHEWBERMAN

GI#1730|@GREGISENBERG

GT#121|@GARRYTAN

220

Andrew Curran argues recursive self-improvement will allow the U.S. government to nationalize AI labs and run them without human staffResearcher Andreas Kirsch warns this eliminates human whistleblowers.

19h|87.8k722219

AC#682|@ANDREWCURRAN_

SK#667|@SEBKRIER

T(#501|@TEORTAXESTEX

AK#271|@BLACKHC

230

ARIA Chair Matt Clifford says the BBC News homepage failed to feature major AI developments in its top 30 storiesResearch engineer fofr says existing coverage underrepresents AI's significance.

18h|110.1k1.5k52

FO#1853|@FOFRAI

MC#1529|@MATTHEWCLIFFORD

RO#1315|@KRISHNANROHIT

240

Nature Medicine study finds general-purpose frontier LLMs outperform specialized clinical AI tools on medical benchmarksGemini 3.1 secured the top consensus ranking among clinicians

1d|31.2k15773

NK#1615|@NEALKHOSLA

BB#1463|@BYERSBLAKE

RP#1257|@ROHANPAUL_AI

ET#816|@ERICTOPOL

TM#613|@ISCIENCELUVR

250

Pete Hegseth says the Department of War permanently evicted Anthropic from its facilities three months agoObservers warn the restriction benefits international competitors like China.

9h|30.1k35219

RR#1394|@RAJKO_RAD

YI#1023|@YISHAN

MS#585|@MARTINSHKRELI

T(#501|@TEORTAXESTEX

AV#350|@ASHLEEVANCE

GM#178|@GARYMARCUS

MB#57|@MILES_BRUNDAGE

260

Google DeepMind finds supervised fine-tuning, not reinforcement learning, is the primary driver of safety behaviors in Gemini modelsThe finding replicates a year-old study by Siva Reddy

12h|21.7k278121

AC#1278|@ARTHURCONMY

SK#667|@SEBKRIER

SR#336|@SIVAREDDYG

NN#254|@NEELNANDA5

270

ErdosBench math benchmark rerun ranks Kimi 2.7 second overall, ahead of GPT-5 and behind Fable 5It also evaluated Qwen 3.7 Max and Grok 4.3.

14h|139.4k1.1k290

ZE#1695|@ZEPHYR_Z9

⿻A#595|@IAMTRASK

AC#484|@ANDREW_N_CARR

280

Anthropic to host Claude Build Day as commentators parody the event with a fictional 'Claude Fable 5' releaseThe actual event offers $150,000 in developer API credits.

1d|73.7k1k44

B(#1623|@ANDERSONBCDEFG

DS#975|@DANSHIPPER

RS#447|@SCOBLEIZER

MB#57|@MILES_BRUNDAGE

290

Garry Tan, Y Combinator President and CEO, argues builders must abandon legacy frameworks for direct, hands-on AI experimentationInvestor James Cham supported the call for active implementation.

11h|96.4k1.1k237

JC#1402|@JAMESCHAM

GT#121|@GARRYTAN

300

Anthropic disputes White House claims that CEO Dario Amodei was unreachable during a "Fable" jailbreak security inquiryThe company acknowledged a 75-minute delay before Amodei joined.

2h|39.1k1.1k146

AV#350|@ASHLEEVANCE