Tessera
Folgen

Matthew Oxley

@matthewoxley.bsky.social

Co-founder of @chiefstudio.co.uk & @theick.app ✌️ Digital nomad in Lanzarote 🇪🇸

340 Posts421 Follower100 Folgt
ali alkhatib@ali-alkhatib.com · last yr.

nope. anthropic made a dataset including fake emails implying a nonexistent engineer was having an affair and the chatbot emitted words related to words from those fake emails implying a fictitious affair. then a bunch of dipshit frauds at anthropic made the home alone face and ran around the house.

Phil Lewis@phillewis.bsky.social · last yr.

An Amazon-backed AI model tried to blackmail engineers who threatened to shut it down, a safety report revealed In tests, Anthropic's Claude Opus 4 would resort to "extremely harmful actions" to preserve its own existence

JD Shadel@jdshadel.com · last yr.

“If they couldn’t be bothered to write this, why should we be bothered to read it?” ‼️ In their cathartic, broligarchy-rattling new book, “The AI Con,” @emilymbender.bsky.social and @alexhanna.bsky.social are handing out needles so you, too, can burst the hype bubble. www.esckey.co/do-you-belie...

Do you believe in hope after “AI” hype? Emily M. Bender and Alex Hanna make the case

In this extended briefing, “TL;DR” = “Totally Legit; Def. Read” because we review “The AI Con,” the spicy new book ready to rattle the broligarchy. Plus, the authors answer ESC KEY .CO’s questions abo...

esckey.co

Emily M. Bender@emilymbender.bsky.social · last yr.

LLMs are nothing more than models of the distribution of the word forms in their training data, with weights modified by post-training to produce somewhat different distributions. Unless your use case requires a model of a distribution of word forms in text, indeed, they suck and aren't useful.

Hank Green@hankgreen.bsky.social · last yr.

There are a lot of critiques of LLMs that I agree with but "they suck and aren't useful" doesn't really hold water. I understand people not using them because of social, economic, and environmental concerns. And I also understand people using them because they can be very useful. Thoughts?

Timnit Gebru@timnitgebru.blacksky.app · last yr.

Here are a few questions the NYT propaganda-as-news-service could have answered instead of the ridiculous ones they asked.

Timnit Gebru@timnitgebru.blacksky.app · last yr.

He could, for instance, have written about any number of topics. Here are some ideas for free. 1. What cults are the Anthropic crew in, and what is the ideological ancestry of these cults? (We've done the work for him here, hint hint, its eugenics: firstmonday.org/ojs/index.ph...).

Sophie Oxley@sophie.chiefstudio.co.uk · last yr.

#Copywriters. You are still needed, my god, you are still needed. But the market that 'gets' the value you bring is completely different from the pre-AI era. It's also a lot more fun to work with. Keen to find out more? Join Writers Against the Algorithm! I'm speaking tomorrow on this subject 💪

Bild
Matthew Oxley@matthewoxley.bsky.social · last yr.

Super proud to get our little agency (it’s just me and @sophieoxley.bsky.social) accepted into @zyppy.com… It’s a super cool and much needed human-curated directory of the best marketing agencies and consultants. Thanks Cyrus 🙏

Chief Studio@chiefstudio.co.uk · last yr.

Woohoo, we’ve just been accepted to Zyppy List! 🕺 Check out our listing and find out a lot more about who we are and what we do… Thanks @zyppy.com zyppy.com/list/compani...

Margaret Mitchell@mmitchell.bsky.social · last yr.

New great piece from @emilymbender.bsky.social explaining the problem with "AI" framing. "If what you are doing is sensible and grounded science, there is...a more precise way to describe it that bolsters rather than undermines the interest and value of your research." buttondown.com/maiht3k/arch...

Calling All Mad Scientists: Reject "AI" as a Framing of Your Work

By Emily tl;dr: Every time you describe your work that involves statistical modeling as "AI" you are lending your power and credibility to Musk and DOGE. Mad...

buttondown.com

Jay Acunzo@jayacunzo.com · last yr.

I wrote about the kind of content marketing that actually works in 2025. What got you here won't get you there. Time to face some hard truths. jayacunzo.com/blog/the-kind-of-content-marketing-that-actually-works-in-2025

The Kind of Content Marketing that Actually Works in 2025

Marketing isn't simple. But understanding it can be. I need that reminder now and again. Maybe you do too. I started my career working for brands like Google as a digital media strategist and two st...

jayacunzo.com

Matthew Oxley@matthewoxley.bsky.social · last yr.

“Change happens by listening and then starting a dialogue with the people who are doing something you don't believe is right.” Gotta love Jane, right? So many good ones in light of today’s shit show in the Oval Office.

Justin Jackson@justinjackson.ca · last yr.

"The greatest danger to our future is apathy. We can't all save the world in a dramatic way, but we can each make our small difference, and together those small differences add up. The question is: What kind of impact do you want to make?" – Jane Goodall

@theick.app · last yr.

Straight from a Google Gemini 2.0 Flash marketing email: Get ready for many actionable insights, a productivity boost, and the seamless experience of… well, doing things seamlessly. 🤦‍♀️ #ai

Screenshot of a Google Gemini marketing email uploaded to the ick app, which highlights overly generic and exaggerated language. Words like 'actionable insights,' 'boost,' and 'seamlessly' are flagged. The app includes different modes like Dyslexia Friendly and Lightbox.