Berkan Ottlik retweeted
I think people think that non-attachment leads to not caring, when in fact it can just lead to continuous motion, which is what actually caring looks like. You can move from moment to moment without sinking into the trap of negative minima or rumination or artless circling.
Berkan Ottlik retweeted
Excited to share that my book, The Worlds I See, will be published on Nov 7, 2023 from @Flatironbooks & @melindagates' Moment of Lift Books. I believe AI can help people & I hope you’ll come along on the journey. Preorder momentoflift.com/the-worlds-… #MoemtnofLiftBooks #WorldsISee 1/
Berkan Ottlik retweeted
Big news: we've agreed to acquire @MosaicML, a leading generative AI platform. I couldn’t be more excited to join forces once the deal closes. databricks.com/mosaic-news
Berkan Ottlik retweeted
Unexpected description of GPT4 architecture from geohotz in a recent interview he gave. At least it’s plausible.
Berkan Ottlik retweeted
OpenAI's Discord is literally a prompt goldmine.
It's where all the prompt nerds hang out and design the craziest prompts.
Here are the 10 best mega prompts I found in there:
Berkan Ottlik retweeted
Adam, a 9-yr old optimizer, is the go-to for training LLMs (eg, GPT-3, OPT, LLAMA).
Introducing Sophia, a new optimizer that is 2x faster than Adam on LLMs. Just a few more lines of code could cut your costs from $2M to $1M (if scaling laws hold).
arxiv.org/abs/2305.14342 🧵⬇️
Berkan Ottlik retweeted
1/ @openai & @AnthropicAI bills got you down? a new paper from Lingjiao Chen, @matei_zaharia, & @james_y_zou last week shows you can use "LLM cascades" to cut down on cost (and even improve accuracy!!)
Berkan Ottlik retweeted
Sunday read.
This is hilarious. Engineers are competing who creates the worst UI on Reddit.
I think the winners are:
1/ Enter your phone number
Berkan Ottlik retweeted
bbe (big builder energy) in nyc today 😃
congrats to all of the super talented hackers & thanks @contrary @betaworks @lux_capital for hosting!
Berkan Ottlik retweeted
@ylecun closing his presentation with some conjectures #phildeeplearning
Berkan Ottlik retweeted
Twitter is an anti-compassion machine that stokes our tendency to dehumanize and even hate people we don't know personally.
Four little exercises I use to combat that and crank up the compassion in my mind:
Berkan Ottlik retweeted
From the amazing @jcrwhittington, with @dorrell_will, @SuryaGanguli, a mathematical justification for the neuron doctrine!! Explains why single neuron responses are often so beautiful in naturalistic (composable) tasks and shows how to get these disentangled reps in deep-nets.
If everything's a big manifold, why do neurons often code for human-interpretable factors? In arxiv.org/abs/2210.01768 we show the most efficient biological representation puts different factors in different neurons. This lets us build machines that disentangle too! 1/9
Berkan Ottlik retweeted
Replying to @sashachapin
basically all the happiest people I know spend like 50% of their life wandering around
Berkan Ottlik retweeted
Replying to @andy_matuschak @mayli
For me, the critical shift was to stop believing that this "not enough" feeling was serving me. I used to think it was the root of my drive - that if I felt I'd done enough, I'd stop doing things. Turns out, now that I think I do enough, I accomplish far, far more.
Berkan Ottlik retweeted
AGI was fringe in the ML world until last year when it went mainstream. I predict AGI safety will likewise become mainstream ML in 2-3 years
Berkan Ottlik retweeted
“The most useful skill you can acquire is the ability to help other people feel safe.”
ava.substack.com/p/safety?s=…



























