[Thread] A new US paper shows the best frontier LLM models achieve 0% on hard real-life Programming Contest problems, domains where expert humans still excel (Rohan Paul/@rohanpaul_ai)

Rohan Paul / @rohanpaul_ai:
[Thread] A new US paper shows the best frontier LLM models achieve 0% on hard real-life Programming Contest problems, domains where expert humans still excel  —  This is really BAD news of LLM's coding skill. ☹️ The best Frontier LLM models achieve 0% on hard real-life Programming Contest problems, domains where expert humans still excel. LiveCodeBench Pro, a benchmark composed of problems from Codeforces, ICPC, and IOI ("International [image]

Tinder unveils Double Date, bringing the double dating feature back to the US after shelving it in 2017 over privacy concerns, rolling out globally in July (Natalie Lung/Bloomberg)

Natalie Lung / Bloomberg:
Tinder unveils Double Date, bringing the double dating feature back to the US after shelving it in 2017 over privacy concerns, rolling out globally in July  —  Tinder is bringing a double-dating feature back to the US, revamping a controversial product it shelved in 2017 over privacy concerns …

Amazon Ads and Roku sign a new ad initiative to give media buyers access to 80%+ of US connected-TV households, set to launch by Q4 2025, using Amazon's DSP (Dade Hayes/Deadline)

Dade Hayes / Deadline:
Amazon Ads and Roku sign a new ad initiative to give media buyers access to 80%+ of US connected-TV households, set to launch by Q4 2025, using Amazon's DSP  —  In a milestone agreement, Amazon Ads and Roku are teaming on a new advertising initiative that will give media buyers access to more than 80% of U.S. connected-TV households.

A survey of UK academic integrity violations finds nearly 7,000 proven AI cheating cases in 2023-24, or 5.1 per 1,000 students, up from 1.6 per 1,000 in 2022-23 (Michael Goodier/The Guardian)

Michael Goodier / The Guardian:
A survey of UK academic integrity violations finds nearly 7,000 proven AI cheating cases in 2023-24, or 5.1 per 1,000 students, up from 1.6 per 1,000 in 2022-23  —  Guardian investigation finds almost 7,000 proven cases of cheating - and experts says these are tip of the iceberg

TikTok rolls out new AI ad features, including the ability for marketers to use text or images to create AI-generated video, as part of its Symphony ad suite (Kurt Wagner/Bloomberg)

Kurt Wagner / Bloomberg:
TikTok rolls out new AI ad features, including the ability for marketers to use text or images to create AI-generated video, as part of its Symphony ad suite  —  TikTok is rolling out a set of new artificial intelligence-powered advertising features, including the ability for marketers …

A look at Apple's upgrades to CarPlay, iPadOS, and visionOS, and the new Liquid Glass redesign of its operating systems announced at WWDC (Mark Gurman/Bloomberg)

Mark Gurman / Bloomberg:
A look at Apple's upgrades to CarPlay, iPadOS, and visionOS, and the new Liquid Glass redesign of its operating systems announced at WWDC  —  Apple's unveiling of iOS 26 and the Liquid Glass interface was the most discussed part of its developers conference, but the real gems were its enhancements …

New York City-based Abacum, which offers AI-powered financial planning and analysis tools to companies, raised a $60B Series B led by Scale Venture Partners (Chris Metinko/Axios)

Chris Metinko / Axios:
New York City-based Abacum, which offers AI-powered financial planning and analysis tools to companies, raised a $60B Series B led by Scale Venture Partners  —  Abacum, business financial planning startup, landed a $60 million Series B round led by Scale Venture Partners, co-founder and CEO Julio Martínez tells Axios Pro.

Dealogic: tech deals accounted for $421B of the $1.67T in global deals announced in the first five months of 2025, with 75% of tech M&A involving AI software (Milana Vinn/Reuters)

Milana Vinn / Reuters:
Dealogic: tech deals accounted for $421B of the $1.67T in global deals announced in the first five months of 2025, with 75% of tech M&A involving AI software  —  Weighed down by tariffs and geopolitical uncertainty, dealmaking has slowed to a crawl across most industries except one: the unglamorous world of data infrastructure.

Anthropic details how it built its multi-agent Claude Research system, claiming significant improvements in internal evaluations over single-agent systems (Anthropic)

Anthropic:
Anthropic details how it built its multi-agent Claude Research system, claiming significant improvements in internal evaluations over single-agent systems  —  Our Research feature uses multiple Claude agents to explore complex topics more effectively.  We share the engineering challenges …

A look at seven rebuttals to Apple's paper on limitations of Large Reasoning Models, and why none make a compelling case (Gary Marcus/Marcus on AI)

Gary Marcus / Marcus on AI:
A look at seven rebuttals to Apple's paper on limitations of Large Reasoning Models, and why none make a compelling case  —  Also: another paper that seals the deal  —  The Apple paper on limitations in the “reasoning” of Large Reasoning Models, which raised challenges for the latest scaling hypothesis, has clearly touched a nerve.

Salt Lake City, which faces a staffing shortfall, plans to implement AI-assisted 911 call triaging to handle ~30% of about 450K non-emergency calls per year (Salt Lake Tribune)

Salt Lake Tribune:
Salt Lake City, which faces a staffing shortfall, plans to implement AI-assisted 911 call triaging to handle ~30% of about 450K non-emergency calls per year  —  Move to artificial intelligence for some nonemergency calls is expected to help human responders focus on the urgent ones.

A profile of Chess.com, which has popularized online chess among pro athletes, as CEO Erik Allebest says the site has over $100M in annual revenue and 35M MAUs (Ira Boudway/Bloomberg)

Ira Boudway / Bloomberg:
A profile of Chess.com, which has popularized online chess among pro athletes, as CEO Erik Allebest says the site has over $100M in annual revenue and 35M MAUs  —  Social media is a minefield for pro sports players.  But online chess offers rankings, instant analysis, wholesome interactions and few trolls.

How Chinese AI companies bypass US chip restrictions by transporting hard drives with AI training data to Malaysian data centers housing advanced Nvidia chips (Wall Street Journal)

Wall Street Journal:
How Chinese AI companies bypass US chip restrictions by transporting hard drives with AI training data to Malaysian data centers housing advanced Nvidia chips  —  Engineers carry data to countries where Nvidia chips are available, frustrating Washington's aims

The Army Reserve launches Detachment 201, an innovation corps with Palantir CTO, Meta CTO, and other tech execs advising on AI and tech as lieutenant colonels (Heather Somerville/Wall Street Journal)

Heather Somerville / Wall Street Journal:
The Army Reserve launches Detachment 201, an innovation corps with Palantir CTO, Meta CTO, and other tech execs advising on AI and tech as lieutenant colonels  —  Silicon Valley executives are joining a new innovation corps in Army Reserve  —  The nerd brigade is reporting for duty.

Google plans to shut down Android's Instant Apps, which let users launch apps without installing them, later this year citing low usage and engagement (Dominic Preston/The Verge)

Dominic Preston / The Verge:
Google plans to shut down Android's Instant Apps, which let users launch apps without installing them, later this year citing low usage and engagement  —  The installation-free apps will join the Google graveyard in December. … Google has confirmed that it plans to shut …