Lifelong Learner✨Building
Brisbane, Australia
Joined May 2012
- Tweets2.2K
- Following433
- Followers291
- Likes23.8K
Melissa Wyborn retweeted
Omega-3s seem to have pharmacological potential similar to that of a drug, but a risk profile more akin to nutrient
Very powerful
For most people, 2g/day (EPA + DHA) increases their omega-3 index from 4% to 8%, which is associated with a 5-year increase in life expectancy
Melissa Wyborn retweeted
Too many of us reach for the "magic pill" before we get the basics right. Before rushing into medications or other interventions, it’s important to maximize our body's natural healing potential.
Modern "healthcare" rarely focuses on these Core Four aspects of wellbeing. Without getting these right, it's hard to justify exposing yourself to a pharmaceutical.
🛌 Sleep: Innumerable studies show that getting 7-9 hours of quality sleep nightly reduces the risk of chronic diseases and lengthens life expectancy. Sleep is the unspoken hack, having vast implications on our endocrine system, muscle regeneration, brain development, and a host of other positive impacts. 😴💤
💪 Exercise: Regular physical activity can not only aid weight management but also enhances overall treatment effectiveness. The benefits to regular exercise are many. We need to tailor our regimens for our own lives, but must importantly ensure we maintain adequate muscle mass to be functional in our later years. 🏃♀️🏋️♂️
🥗 Nutrition: A colleague once joked that Americans have the most expensive pee in the world because of all the supplements we take... Think of your nutrition as the ultimate, three times a day supplement that your body is designed to metabolize! (not that I'm against supps, but we should strive to optimize our nutritional biochemistry through real whole foods first).🍎🥦
🧠 Relationships/Mental Well-being: Nurturing meaningful connections and prioritizing mental health is the foundation of any health improvement. Social isolation carries > risk than smoking a pack of cigarettes/day! Surround yourselves with people you love and who make you smile. ❤️🧠
Without dialing in on all these aspects of your life first, it is almost pointless to seek pharmaceutical or medical interventions. That's why we believe it is so important to have an entire TEAM in your corner helping to get you optimized across all aspects of your life.
Join me as I do a #DEEPdive into Dr. Deeps Core 4 over the next four days 🌟💚 #WellnessFirst #HolisticHealth #EmpowerYourself #HealingFromWithin
@rebelhealth_ @jackmurphylive @AbudBakri
Melissa Wyborn retweeted
I'm most alive when I'm:
• Creating (ideally something that can be seen, touch, felt)
• Moving (lifting, bouldering, cycling, etc.)
• Connecting (authentically with others)
Perhaps this is universal.
How to get into bigger rooms? This is an excellent response! 🔥
He didn't answer the question of getting into bigger rooms.
How you do that? Let me help Ted.
You do something with excellence...and at first you will be doing that thing with people who are *not* in particularly big rooms...but then, they have friends..maybe some friends who are up a bit further than they are. They will say, "Oh I know a guy...Ted...he's phenomenal" and you get referred to them. And then you do excellent work again. And then, that person has a friend - maybe this time - in larger circles - someone who needs something that you can do, and they refer you in because you can do it. You do excellent work yet again. You hold your standard. You go above and beyond. And you keep doing the solid work, keep improving yourself, and keep scaling, and keep putting yourself in position to be in those larger rooms. You are consistent, you are reliable, you are excellent. You both strive to be the best at whatever you do - and you intentionally start to be where people are who could benefit from that excellence. You now have a history of execution that inspires trust and can be trusted to execute at the highest levels - you solve problems reliably and are top tier. At this point, you start a flywheel and eventually you will get into almost any room you want. Everyone who needs that thing will want you - because now you're not just "Ted" - you are "the only" possible person someone might want. Your reputation gets you in that room before you ever set foot in it.
AI + Productivity + Education: Personal study coach that uses learning styles and science backed productivity methods to create customized learning journeys. Layer storytelling on top and boom! 🔥
"The last big platform shift, mobile, gave rise to companies like Airbnb, Instagram, and Uber. We expect to see the same with AI." 🔥
Introducing the Abundance Agenda.
We believe AI will turn luxuries into commodities - making consumers happier, healthier, and more productive than ever before.
And a new generation of AI-native companies will lead the way.
More from @a16z consumer 👇
Just generated a draft pitch deck in seconds, how fun! 😎 @gammaapp
3⃣ Intelligent editors that reduce work + boost creativity.
Editing has historically been a massive frustration as it requires detailed, manual work.
Just like AI allows for natural language generation, we expect to see this for editing - check out @gammaapp's AI assistant.
“Hair salons have been around forever. Alli decided to focus on 1 part of the experience & make it the absolute best.” 🔥
Fantastic insights!🔥 My favourite is around education and entertainment 👀
🥁 DRUMROLL PLEASE!! 🥁
Having synthesized feedback from 250+ Gen Zers, we at Redpoint could not be more excited to spill *THE TEA ON GEN Z* ☕️☕️☕️
redpoint.com/start/written/t…
“…reminder that a well-placed "f*** it" could be all that stands between you and the start of an even happier life.”🔥
I arrived in London 3 years ago to start a job I didn't care about.
0 direction, 0 long-term plans. Happy enough, but aimless.
Today, I'm leaving the city with a life that is almost completely unrecognisable...
Sometimes, the entire direction of your life rests on your willingness to say "f*** it" at just the right time...
Here's the story:
1 year after arriving in London, @AliAbdaal put out job applications.
I'd only discovered his channel a few weeks before, but I applied on a whim.
Actually, I applied for 3 different roles.
"F*** it", I thought. "I like YouTube. I can write. I once had a gaming channel with 1500 subscribers. F*** it."
Despite having no experience, I had a sense that "working for a YouTuber" was something that would suit me. I literally didn't care what role, I just wanted one.
As the application process drew on, I spent more and more time imagining what this "new life" might look like, and the reality of my current job became more and more stark.
"What if, rather than doing 12-hour night shifts... I got to write YouTube videos for a living?"
Hilariously, as I started writing this tweet, I actually found some old diary entries from the time I was waiting to hear if I got the job or not.
Apparently, I titled it:
"The Gradual Breakdown of a man who hated his job" 😂
Evidently, I was keen to leave. The allure of YouTube was more powerful every day.
But, as time went on, I started to convince myself that a job like "working for a YouTuber" was simply not something I'd ever end up doing.
Like I said, I had 0 experience. It sounded too fun. Life didn't work like that.
Then, inexplicably, 2 months later...
I got it.
I handed in my notice, and my day-to-day life did a complete 180.
Night-shifts were now "flexible hours".
Limited annual leave was now "as much as you want, just don't take the p***"
Data entry was now "writing scripts for YouTube".
I woke up feeling grateful almost every day.
But even that version of me wouldn't have believed how much would change again in the following year.
Fast-forward to 2023, and I have a community of nearly 10,000 people across my platforms, just from "writing about writing".
I get to choose my clients based on who I have the most fun working for.
I've gotten to know incredibly talented people from across the world.
I'm building an actual business, wtf.
And, most excitingly, I'm leaving London with my girlfriend (who I met just after moving here) to go live and work in Europe for a while.
Because f*** it, why not?
So take this as a reminder that a well-placed "f*** it" could be all that stands between you and the start of an even happier life.
Of course, there's a little bit of luck in every bit of fortune. But you've gotta be willing to grab it when it comes.
I'm endlessly thankful to 2021-George for saying "f*** it" at just the right time ❤️
Melissa Wyborn retweeted
More than 2B users now watch YouTube Shorts monthly - exceeding TikTok's 1.2B MAUs 🤯
And YouTube is doubling down w/ new creator features:
- Livestreaming
- Split screen "collabs" (TikTok duets)
- Repurpose long form vids -> shorts
- Auto-import sound / effect combo
Melissa Wyborn retweeted
The real genius behind Barbie's success wasn't the $150 million marketing budget.
Prior to the film's launch, Mattel (owner of Barbie) pushed their licensing agreements to 100+ brands.
Its why they felt *everywhere*
Brands that collaborated include:
- AirBnB (rent Ken's dream house)
- Burger King (pink burger, ew)
- Xbox (custom hardware)
Those agreements typically look like this...
1. Small brand asks big brand to collab
2. Big brand requests minimum royalty
3. Small brand struggles to hit that #
And then proceeds to complain about it.
For the movie release, Mattel did the opposite.
They actively pursued brands.
Lowered their minimum royalty figure
Plus, streamlined the product approval process.
It worked great. But isn't a "obvious" move.
There's precendent for the opposite.
Hasbro, another toy conglomerate, has been restrictive with the licensing of their major franchises, like:
→ Transformers
→ Dungeons & Dragons
→ G.I. Joe
Meaning they fully control the marketing, the films, and the products sold post release.
To do this, they bought a production studio for $3.8 billion in 2019.
Last year, they sold it for $500 million.
Yikes.
Moral of the story?
Crowdsourcing you're marketing works.
Barbie gave other companies' marketing teams an easy opportunity.
Leverage us to create a unique drop.
It can be pink, fun, trendy, you name it.
Just get people to talk about the movie.
Doing so helped them reach billions of consumers with other people's licensed products.
...and the brands paid them to do it. – at Austin, TX
Melissa Wyborn retweeted
Building at the intersection of games x healthcare?
Normies think games are toxic for society. But we believe games are actually the key to solving many consumer/population health problems.
Why I'm excited for games x healthcare, and why you should apply for SPEEDRUN if you're building in this space🧵
speedrun.a16z.com
Melissa Wyborn retweeted
we've been working with a lot of accelerators lately, so our team put together this doc with the 100 top accelerators including application links, deadlines, and other criteria to make the search easier for founders thinking about joining an accelerator. access here:
Melissa Wyborn retweeted
The best opening scenes in movie history — and why:
1/ The Dark Knight
No backstory. Just an insane villain executing a brilliant plan.
Batman needs the Joker, not the other way around.
Melissa Wyborn retweeted
Calling it now: interactive fan fiction is going to be a killer use case of QLoRA.
It’s now feasible for consumers to fine-tune an LLM on the massive amounts of fan fiction that’s available publicly.
I expect we’ll see models for individual worlds and even specific characters.
I can't believe I've just fine-tuned a 33B-parameter LLM on Google Colab in a few hours.😱
Insane announcement for any of you using open-source LLMs on normal GPUs! 🤯
A new paper has been released, QLoRA, which is nothing short of game-changing for the ability to train and fine-tune LLMs on consumers' GPUs.
In a few words:
QLoRA reduces the memory usage of LLM fine-tuning without any performance tradeoffs compared to standard 16-bit model fine-tuning.
This method enables 33B model fine-tuning on a single 24GB GPU and 65B model fine-tuning on a single 46GB GPU. This is incredible! 😍
More specifically, QLoRA uses 4-bit quantization to compress a pre-trained language model. The LM parameters are then frozen, and a relatively small number of trainable parameters are added to the model in the form of Low-Rank Adapters.
During finetuning, QLoRA backpropagates gradients through the frozen 4-bit quantized pretrained language model into the Low-Rank Adapters. The LoRA layers are the only parameters being updated during training. Read more about LoRA in the original LoRA paper (arxiv.org/abs/2106.09685). 🤓
QLoRA has one storage data type (usually 4-bit NormalFloat) for the base model weights and a computation data type (16-bit BrainFloat) used to perform computations. QLoRA dequantizes weights from the storage data type to the computation data type to perform the forward and backward passes, but only computes weight gradients for the LoRA parameters, which use 16-bit bfloat. The weights are decompressed only when they are needed, therefore the memory usage stays low during training and inference. Beautiful!😱
QLoRA tuning is shown to match 16-bit finetuning methods in a wide range of experiments. In addition, the Guanaco models, which use QLoRA finetuning for LLaMA models on the OpenAssistant dataset (OASST1), are state-of-the-art chatbot systems and are close to ChatGPT on the Vicuna benchmark. This is an additional demonstration of the power of QLoRA tuning.
Their Guanaco models are reaching 99.3% of the performance level of ChatGPT while only requiring 24 hours of fine-tuning on a single GPU. You can actually do it in Google Colab.
📚 Links-
QLoRA Paper - arxiv.org/pdf/2305.14314.pdf
Colab for inference - colab.research.google.com/dr…
Colab for fine-tuning - colab.research.google.com/dr…
GitHub Repository- github.com/artidoro/qlora
Use it with HuggingFace - huggingface.co/blog/4bit-tra…
