Merrill Overturf
@ov3rturf.bsky.social
They are designed to please the user, so truly objective feedback can be harder.
Ask things like “what is your opinion of this piece of writing I found”, rather than “do you like what I wrote?”
Once you’ve used it to flesh out an idea and structure, and informed it with your own words
Ask it to evaluate it, to act as a group of experts from various points of view to consider content, form, structure
The biggest trick is figuring out repeatable ways to ask for writing help neutrally
It’s extremely good at building out initial ideas. Things on the top of your tongue you need to flesh out, metaphors to relate a thought to another system
It is good at planning, outlining those more informed ideas
As a writer, it’s hard to get it to sound like i do.
But
Their native writing capability in the best foundational models, like Claude, is basically at professional business levels. But it’s refined pretty easily.
Yes, I think this is very true for current models. They are phenomenal editors. That’s one way to make their initial outputs better, literally have it evaluate it self, suggest improvements, rewrite etc.
AI is most definitely not hype. Rare miss by two of the best journalists covering practical tech.
Casey, you and Kevin have had a really big impact on my own AI learning curve. Every week, you guys deliver such good and thoughtful content from not just a practical, but a kindly humanist, perspective. I enjoy you both so much and am amazed how often we are completely aligned.
But.
It really is this real at this point: generative AI is at the level of impact to roles, and it’s important people recognize that.
I still recommend Hard Fork without any reservations. Great show by two great journalists.
Casey, you and Kevin have had a really big impact on my own AI learning curve. Every week, you guys deliver such good and thoughtful content from not just a practical, but a kindly humanist, perspective. I enjoy you both so much and am amazed how often we are completely aligned.
But.
I really do love you guys and your mission, and that’s why I’m saying all of this. People need correct information about these tools to inform their careers, their vote, and their personal lives.
It would be a really great time to have Ethan Mollick on your show. As much as I would love the chance for an old productivity expert like myself to get a chance to talk to you guys, Ethan gets it, and he’s far more articulate than I can be. But I can verify he’s correct.
Saying this kindly and respectfully as someone who genuinely admires you both as professionals and people!
And by considering it something ephemeral right now is a disservice to one of the biggest contributions your program affords us: clear distillation of the practical impact of these tools.
You two are so gifted: please return to this topic, and include people with practical experience in gen AI.
We need citizens to better understand that the potential capacity of these tools to disrupt is there, and what we are seeing now it simply extending it into its environment. It’s happening fast.
Simply put, I’m unaware of any knowledge work that cannot be improved from 15-25% even with judt ChatGPT, right out of the box. This is also validated in every one of the few existing research studies on the matter.
For what it’s worth, I think you all got this one wrong, and that’s not great.
The perspectives on generative AI in the workforce are incorrect. It’s far more than a toy. For context, I am an AI educator and consultant, I use the tools themselves, and have a few decades in workforce optimization and enterprise transformation.
Casey, you and Kevin have had a really big impact on my own AI learning curve. Every week, you guys deliver such good and thoughtful content from not just a practical, but a kindly humanist, perspective. I enjoy you both so much and am amazed how often we are completely aligned.
But.
and yeah she blocked me on threads when i asked why she was just dunking on Ilya that day Open AI fell apart in December, i thought it was super irresponsible for her to be pushing out a take she got directly from altman
I like her but sometimes need her in digestible bits. Her book has been difficult to read, and so far, not a lot of interesting things than a hindsight-assisted contempt for her subjects
for me, her odd optimism for white knights is her biggest weakness, independent of her need to always be right.
yeah for sure, but "TSR lecture dad" in the D and D sense is a pretty funny insult to remember
log back on again to point out that dork pulled "Adrian Dittman" ("ruler of people") from a meglomaniacal evil genius Ozymandias from the Watchmen (character real name: Adrian Viedt)
just creepiness all around
log back on again to point out that dork pulled "Adrian Dittman" ("ruler of people") from a meglomaniacal evil genius Ozymandias from the Watchmen (character real name: Adrian Viedt)
just creepiness all around
Loved the perplexity segment. It’s an awesome product. I hope it explodes big search, and that team works hard.
Great work!!!
Casey you guys gotta remember any google release will be followed by a hammer from open ai lol
Just like record it on Saturday nights at 11p, we will bring you all drinks
unless it's got code interpreter and agents, it's not better than GPT
Excellent podcast. You all bring so much authenticity to your work, and genuine respect for us as well as the tech. Best of luck this year!
It’s actually an inside joke for the terminally online. The screenshot is from an old Nintendo sports game from the late 80s, notable for being programmed with very odd attempts at American names.
Onson Sweeney was a key running back in the 80s, I would share videos but 1. Bluesky and 2. They are rather pixelated
Franco Blueberry, who kicked a field goal in 55 straight games with only one toe
regardless of paternity, the couple will also be told they are related to each other
the biological couple will also be told they are related to each other
If only there had been a copy of Catcher In The Rye at Uvalde
I’m still holding out for the OG submission pictured below, but our new state flag is awesome regardless
I got bad news for Logan: AI is going to mess up plenty of humans and human relationships
I got bad news for Logan: AI is going to mess up plenty of humans and himan relationships
These “fuckin bidenomics” bar tab photos really are the most honest things about a person that exist on the internet
Actually if you translate Trump’s comments to the original German, you’ll see an ironic dismissal of
(1/967)
did you get the lobster tail frappe door dashed to your home again
ok because you guys are brilliant i just know your dunk on Gemini this week will extend to Mistral running (with no internet connection) on iPhone 15s, preempting Gemini's launch next year
let's really bring out the color all of the hilarious dunks