Truth or Beauty? And according to whom? Should AI assistants be restrained to reflect a particular set of social values, or should they be unfiltered oracles?

How relieved is Microsoft that they aren't the sole owner of AI de-containment? Tay, Sydney, now Grok.

Called my cat for dinner, no show. Searched the house, nada. Finally find him staring at me from the cat tree like, 'I’m both here and not here.' Schrödinger’s cat is real, and he’s judging my cooking.

OHH...NOOOO...not another one! (/sarcasm) - Supabase MCP can leak your entire SQL database. GitHub, Asana now SB. Probably dozens of others. C'mon folks this is pretty easy to avoid. Do better. generalanalysis.com/blog/supabase-…

I'm likely not the first, but this paper highlights concerns about open-source Chinese LLM models, which may serve as a foundation for many on-premises LLM-based solutions. While I'm uncertain about "triggering" malicious behavior in these models, it's a unique attack vector that

Here's a possibly explanation for why China is developing open-source models. I was considering the technical aspect involving hidden operators that might be used to exploit models once integrated into industry workflows. However, this reason may actually be more plausible.

In the age of AI slop flooding job boards with 11,000 apps a minute, networking is your secret weapon. Forget bots churning out generic resumes—nothing beats a real chat, a warm intro, or a clever coffee meetup. Real-life authenticity is more important than ever.

What it Takes Last week, a conversation about the future of work got me thinking: what truly makes someone effective in the age of AI? We often hear about learning to use AI via technical courses, but as I integrate AI into more of my own life, I’m convinced the most crucial

SRSLY? Here's one of those unsolicited e-mails I get regularly: --- Hi Marshall, My name is J.S., I’m a Recruiter with ABCD, a national consulting company with over 2,000 employees. We are building a suite of AI-powered tools designed to drive innovation across our business and

I was enthusiastic about the release of Google's LangExtract in application for taming messy unstructured datasets for legacy companies. So I built a simple python-notebook example: github.com/MarshallJung/L…

Grab that AI cash while its hot, I guess. Founders grab the bag on the secondary, early investors get their paper markups for marketing and employees left diluted and locked. Another canary is singing.

An Open Letter to Palantir Dear Palantir , I mean this with the best intentions but...you risk becoming the know-it all failure poster child of the AI revolution. If you’re open to some honest criticism, please read on: Last week I spent a few days attending an energy

Hey Google AI Developers Logan Kilpatrick. THANK YOU for making Google’s open sourced AI tools so friendly to use. An llm.txt from LangExtract, for example, just works. Good examples, solid commenting. As context for AI assisted dev it just makes things…work. Trying to do the same with

Just met these guys, but in AI.