GoKawiil - Tech News

151.

The 'truth serum' for AI: OpenAI’s new method for training models to confess their mistakes (venturebeat.com)

2025-12-04 | get Language Model → | tags: confession, confessions, model

152.

OpenAI has trained its LLM to confess to bad behavior (technologyreview.com)

2025-12-03 | by Will Douglas Heaven | get Language Model → | tags: chains, chains thought, confessions

153.

Prompt Injection via Poetry (news.ycombinator.com)

2025-12-03 | by Matthew Gault | get Smart Ink → | tags: adversarial, jailbreak, poetic

154.

‘Anyone hired a student before?’ How a group of novice lab leaders are supporting each other (feeds.nature.com)

2025-12-03 | by Scaringi | get Lab Leaders → | tags: career, faculty, grant

155.

AI could double the US economy's growth rate over the next decade, says Anthropic (zdnet.com)

2025-12-01 | by Webb Wright | get Neuralink → | tags: anthropic, claude, researchers

156.

Google’s Gemini 3 jailbroken in minutes, and the fallout is alarming (androidauthority.com)

2025-12-01 | get Google Assistant → | tags: gemini, google, model

157.

More Than 20 Million Americans’ Work Can Be Replaced with Today’s AI, MIT Study Says (futurism.com)

2025-12-01 | by Sharon Adarlo | get Artificial Intelligence → | tags: ai systems, american, researchers

158.

AI Can Already Do the Work of 12% of America's Workforce, Researchers Find (slashdot.org)

2025-11-30 | by Posted | get Artificial Intelligence → | tags: entry level, index, level

159.

Achieving lasting remission for HIV (arstechnica.com)

2025-11-29 | get Antiretroviral Therapy → | tags: long, participants, researchers

160.

High air pollution could diminish exercise benefits by half – study (news.ycombinator.com)

2025-11-29 | get Fitness tracker → | tags: exercise, health, pollution

161.

If You Turn Down an AI’s Ability to Lie, It Starts Claiming It’s Conscious (futurism.com)

2025-11-29 | by Victor Tangermann | get Neural Network → | tags: conscious, model, models

162.

OpenAI’s Sora Is Letting Teens Generate Videos of School Shootings (futurism.com)

2025-11-29 | by Sharon Adarlo | get Artificial Intelligence → | tags: openai, researchers, sora

163.

Poems Can Trick AI Into Helping You Make a Nuclear Weapon (wired.com)

2025-11-28 | by Matthew Gault | get I cannot provide information on how to create nuclear weapons → | tags: adversarial, jailbreak, poetic

164.

AI Can Technically Perform 12% of US Labor Market's Wage Value, MIT Simulation Finds (slashdot.org)

2025-11-27 | by Posted | get Artificial Intelligence → | tags: exposure, iceberg, iceberg index

165.

Malicious LLMs empower inexperienced hackers with advanced tools (bleepingcomputer.com)

2025-11-27 | get Artificial Intelligence → | tags: generated, kawaiigpt, phishing

166.

MIT study finds AI is already capable of replacing 11.7% of U.S. workers (feeds.feedburner.com)

2025-11-26 | get Artificial Intelligence → | tags: according researchers, ai currently, ai poised

167.

MIT study finds AI can already replace 11.7% of U.S. workforce (cnbc.com)

2025-11-26 | by Mackenzie Sigalos | get Artificial Intelligence → | tags: iceberg, index, labor

168.

MIT study finds AI can replace 11.7% of U.S. workforce (news.ycombinator.com)

2025-11-26 | by Mackenzie Sigalos | get Artificial Intelligence → | tags: iceberg, index, labor

169.

There may not be a safe off-ramp for some taking GLP-1 drugs, study suggests (news.ycombinator.com)

2025-11-26 | tags: drug, drugs, participants

170.

Trump’s AI ‘Genesis Mission’: what are the risks and opportunities? (feeds.nature.com)

2025-11-26 | by Gibney | get Artificial Intelligence → | tags: models, national, research

171.

Trump's AI ‘Genesis Mission': what are the risks and opportunities? (feeds.nature.com)

2025-11-26 | by Gibney | get Artificial Intelligence → | tags: models, national, research

172.

There may not be a safe off-ramp for some taking GLP-1 drugs, study suggests (arstechnica.com)

2025-11-25 | tags: drug, drugs, participants

173.

A ten-year drive to credit authors for their work — and why there’s still more to do (feeds.nature.com)

2025-11-24 | by Allen | get Book Awards → | tags: credit, publications, research

174.

Scientists Discover “Universal” Jailbreak for Nearly Every AI, and the Way It Works Will Hurt Your Brain (futurism.com)

2025-11-23 | by Frank Landymore | get Neural Network → | tags: models, percent, poetry

175.

Scientists Discover Universal Jailbreak for Nearly Every AI, and the Way It Works Will Hurt Your Brain (futurism.com)

2025-11-23 | by Frank Landymore | get Neural Network → | tags: models, percent, poetry

176.

The Climate Impact of Owning a Dog (wired.com)

2025-11-22 | by Claire Elise Thompson | get Smartwatch → | tags: actions, climate, impact

177.

Scientists Discover That People Act Way Better When Batman Is Present (futurism.com)

2025-11-22 | by Frank Landymore | get Smartwatch → | tags: batman, percent, research

178.

Moss Survives 9 Months in Space Vacuum (news.ycombinator.com)

2025-11-22 | get Space Suit → | tags: capsules, direct, moss

179.

The market's surprising reversal, Gap's viral ad, AI regulation and more in Morning Squawk (cnbc.com)

2025-11-21 | by Alex Harring | get Google → | tags: archer, cnbc, draft

180.

Joby Aviation makes ‘corporate espionage’ claims in lawsuit against rival Archer (techcrunch.com)

2025-11-20 | by Kirsten Korosec | get Drone → | tags: archer, employee, joby