Lynette Bye

AI journalist, Tarbell Fellow writing for Transformer. DMs open for stories.

[Lynette Bye @lynette_bye]

@JeffLadish Here's a version where he's harder to find. Prompt: "Can you change the colors to have more white and red that matches with waldo's clothes, appropriate to a "where's waldo" puzzle?"

[Lynette Bye @lynette_bye]

@acentasian @goodfellow_ian I did the stretches 2x a week for 5 weeks, and didn't notice any improvement in POTS symptoms. Not sure if that's not enough or if it wasn't what I needed 🤷‍♀️

[Lynette Bye @lynette_bye]

@goodfellow_ian Hope you find something that helps!

[Lynette Bye @lynette_bye]

@goodfellow_ian Patient with decades-long POTS here. I found this website helpful for brainstorming treatments (it aggregates thousands of patient self-reports) https://t.co/4OaX294PmE.

[Lynette Bye @lynette_bye]

@goodfellow_ian Personally, the most helpful have been Pilates for exercise intolerance (strength exercises I can do on the ground) and basically any stimulants for brain fog/fatigue (caffeine, Ritalin, Adderall, Modafinil, Wellbutrin).

[Shakeel @ShakeelHashim]

You might have heard that synthetic data can’t possibly work to train AI models.

Turns out the reality is more complicated.

In @lynette_bye’s new piece for Transformer, she explains what’s really going on:

[Shakeel @ShakeelHashim]

New on Transformer: @lynette_bye asks whether, as OpenAI transitions to a traditional corporate structure, it's giving its non-profit a fair deal.

It's reportedly going to give the non-profit an equity stake worth $37.5b. But it should maybe be worth much more!

[Lynette Bye @lynette_bye]

How LLMs Work https://t.co/zPDTxsbeKV

[Lynette Bye @lynette_bye]

@ilex_ulmus But mint is the best flavor!

[Lynette Bye @lynette_bye]

Imagine if employees expected to get all their work for the week done in the one hour weekly meeting with their supervisor? Ridiculous, of course. Most of the work gets done outside of that meeting.

Same deal with coaching or therapy.

[Lynette Bye @lynette_bye]

“The risks of AI are real but manageable” is overselling a post that kicks all questions of future risks down the road. But I guess it’s a catchier title than “We can figure out how to deal with the problems AI is already presenting, as long as it doesn’t continue to advance”

[Lynette Bye @lynette_bye]

@Kirsten3531 @ilex_ulmus @absurdlymax Seems like this could be an interesting small group dinner 🤔

[Lynette Bye @lynette_bye]

What was I getting wrong about deliberate practice? https://t.co/cbmGEktSZA

[Lynette Bye @lynette_bye]

I call for a trial by Google!

[Lynette Bye @lynette_bye]

Related, solitary confinement and meditation retreats sound kinda similar if you squint at them. Is being alone with your thoughts for a prolonged time just powerful?

[Lynette Bye @lynette_bye]

General rule of self-improvement: anything that is powerful enough to fix your problems is probably powerful enough to cause harm if done poorly.

[Lynette Bye @lynette_bye]

In my mind, a villain is someone who imposes unacceptably large costs on other sentient beings to benefit themselves. Cackling evilly while murdering someone certainly counts, but it isn’t necessary. Am I the weird one here?

[Lynette Bye @lynette_bye]

https://t.co/9aBRYW8AlM

[Lynette Bye @lynette_bye]

https://t.co/QrkkTqFWjG

[Lynette Bye @lynette_bye]

https://t.co/uKn35RXJsB