Crooked Bough News Round-up

— MCP: The new “USB-C for AI” that’s bringing fierce rivals together

 

at 4/1/25 5:30am

Model context protocol standardizes how AI uses data sources, supported by OpenAI and Anthropic.

[Category: AI, Biz & IT, Anthropic, API, ChatGPT, chatgtp, large language models, machine learning, MCP, Model Context Protocol, openai]

— What could possibly go wrong? DOGE to rapidly rebuild Social Security codebase.

 

at 3/29/25 8:08am

A safe and proper rewrite should take years not months.

[Category: Biz & IT, Policy, cobol, doge, Elon Musk, Java, syndication]

— Beyond RGB: A new image file format efficiently stores invisible light data

 

at 3/28/25 4:11pm

New Spectral JPEG XL compression reduces file sizes, making spectral imaging more practical.

[Category: Biz & IT, Science, compression, data science, image compression, image formats, JPEG, jpeg xl, multispectrum images, spectral imaging]

— Oracle has reportedly suffered 2 separate breaches exposing thousands of customers‘ PII

 

at 3/28/25 1:41pm

Alleged breaches affect Oracle Cloud and Oracle Health.

[Category: Biz & IT, Security, Data breaches, oracle, oracle health]

— Gemini hackers can deliver more potent attacks with a helping hand from… Gemini

 

at 3/28/25 5:00am

Hacking LLMs has always been more art than science. A new attack on Gemini could change that.

[Category: AI, Biz & IT, Features, Security, Artificial Intelligence, fun-tuning, gemini, google, large language models, LLMs, prompt injections]

— OpenAI’s new AI image generator is potent and bound to provoke

 

at 3/27/25 5:15am

The visual apocalypse is probably nigh, but perhaps seeing was never believing.

[Category: AI, Biz & IT, 4o Image Generation, AI image generator, autoregressive, autoregressive image generator, ChatGPT, chatgtp, dall-e, DALL-E 3, GPT-4o, image synthesis, machine learning, multimodal, multimodal AI, multimodal image generation, openai]

— Broadcom’s VMware says Siemens pirated “thousands” of copies of its software

 

at 3/26/25 8:54am

VMware claims Siemens showed it a list of the VMware products it's using unlicensed.

[Category: Biz & IT, Broadcom, Siemens, vmware]

— Open source devs say AI crawlers dominate traffic, forcing blocks on entire countries

 

at 3/25/25 3:36pm

AI bots hungry for data are taking down FOSS sites by accident, but humans are fighting back.

[Category: AI, Biz & IT, DDoS, machine learning]

— Europe is looking for alternatives to US cloud providers

 

at 3/25/25 7:12am

Some European cloud companies have seen an increase in business.

[Category: Biz & IT, Policy, AWS, azure, cloud, EU, europe, syndication, Trump]

— You can now download the source code that sparked the AI boom

 

at 3/24/25 4:14pm

CHM releases code for 2012 AlexNet breakthrough that proved "deep learning" could work.

[Category: AI, Biz & IT]

— Cloudflare turns AI against itself with endless maze of irrelevant facts

 

at 3/21/25 3:14pm

New approach punishes AI companies that ignore "no crawl" directives.

[Category: AI, Biz & IT, AI security, cloudflare, large language models, machine learning]

— Anthropic’s new AI search feature digs through the web for answers

 

at 3/21/25 1:08pm

Anthropic Claude just caught up with a ChatGPT feature from 2023—but will it be accurate?

[Category: AI, Biz & IT, AI assistants, ai search, AI web search, Anthropic, Anthropic Claude, brave browser, Brave search, ChatGPT, chatgtp, Claude, large language models, machine learning, Simon Willison, web search]

— Study finds AI-generated meme captions funnier than human ones on average

 

at 3/19/25 4:12pm

Mollick proclaims "the meme Turing Test has been passed," but a new study offers a key caveat.

[Category: AI, Biz & IT, Culture, AI study, Ethan Mollick, GPT-4o, Humor, Internet memes, large language models, machine learning, openai]

— Nvidia announces DGX desktop “personal AI supercomputers”

 

at 3/18/25 3:19pm

Asus, Dell, HP, and others to produce powerful desktop machines that run AI models locally.

[Category: AI, Biz & IT, Tech, AI GPU, ASUS, dell, desktop AI PC, desktop PC, DGX, DGX Spark, DGX Station, GPUs, HP, Jensen Huang, Lenovo, machine learning, NVIDIA, Nvidia DGX, Nvidia GTC, personal AI supercomputer, supermicro]

— Nvidia announces “Rubin Ultra” and “Feynman” AI chips for 2027 and 2028

 

at 3/18/25 3:10pm

CEO Jensen Huang says new chips will power robots and billions of AI agents.

[Category: AI, Biz & IT, Tech, AI infrastructure, Blackwell, cloud computing, datacenters, GPUs, IT, Jensen Huang, machine learning, NVIDIA, Nvidia Feynman, Nvidia Vera Rubin, Rubin Ultra, Vera Rubin]

— Farewell Photoshop? Google’s new AI lets you edit images by asking.

 

at 3/18/25 5:15am

New experimental AI allows no-skill photo editing, including removing watermarks. But it's not perfect.

[Category: AI, Biz & IT, Google, AI image generators, ChatGPT, chatgtp, gemini, Gemini 2.0, Gemini 2.0 Flash, google, Google Gemini, image synthesis, machine learning, multimodal AI]

— Large enterprises scramble after supply-chain attack spills their secrets

 

at 3/16/25 8:24pm

tj-actions/changed-files corrupted to run credential-stealing memory scraper.

[Category: Biz & IT, Security, Open source software, supply chain attack, tj-actions]

— Researchers astonished by tool’s apparent success at revealing AI’s “hidden objectives”

 

at 3/14/25 2:03pm

Anthropic trains custom AI to hide objectives, but different "personas" spill their secrets.

[Category: AI, Biz & IT, AI alignment, AI deception, AI research, Alignment research, Anthropic, ChatGPT, chatgtp, Claude, Claude 3.5 Haiku, large language models, machine learning]

— AI search engines cite incorrect news sources at an alarming 60% rate, study says

 

at 3/13/25 3:16pm

CJR study shows AI search services misinform users and ignore publisher exclusion requests.

[Category: AI, Biz & IT, AI confabulations, ai hallucinations, ai search, ChatGPT, confabulations, deepseek, DeepSeek Search, google, Google Gemini, grok, Grok 3, machine learning, microsoft, Microsoft Copilot, openai, Perplexity, xAI]

— AI coding assistant refuses to write code, tells user to learn programming instead

 

at 3/13/25 9:43am

Cursor AI tells user, "I cannot generate code for you, as that would be completing your work."

[Category: AI, Biz & IT, Tech, AI coding, AI paternalism, AI refusals, ChatGPT, chatgtp, cursor, GitHub, large language models, machine learning, Programming, stack overflow]

— Anthropic CEO floats idea of giving AI a “quit job” button, sparking skepticism

 

at 3/13/25 5:15am

"Probably the craziest thing I've said so far," he admitted during an interview.

[Category: AI, Biz & IT, Tech, AI consiousness, AI ethics, Anthropic, Claude, Dario Amodei, Kyle Fish, large language models, machine learning]

As of 4/2/25 1:32am. Last new 3/29/25 10:44am.