Teague Lasser — e/acc retweeted
You can now start building with Gemini 1.5 Flash and Pro models using our API pay-as-you-go service for developers. 🛠️
Flash is designed to be fast and efficient to serve - and we’re increasing the rate limit to 1000 requests per minute.
Find out more → dpmd.ai/3WYse2p
Teague Lasser — e/acc retweeted
Leaning in to the analogy, a computer without an OS isn't going to be very useful. Brains have a bootstrapper that starts the Consciousness OS, but we don't know how it works. An intelligent machine isn't going to become conscious without bootstapping; we need to write the OS.
Teague Lasser — e/acc retweeted
Project Astra is a prototype from @GoogleDeepMind exploring how a universal AI agent can be truly helpful in everyday life. Watch our prototype in action in two parts, each captured in a single take, in real time ↓ #GoogleIO
Teague Lasser — e/acc retweeted
When I was in school the teacher still was drawing on transparencies on an overhead projector. This worked so well because:
1. You could sit while doing it
2. You could erase just as well as a whiteboard
3. You had physical notes after you were done.
Why did we stop???
Teague Lasser — e/acc retweeted
If >50% take the white pill, everything is going to be alright
If you take the black pill... All of the white pill takers are still gonna figure everything out and it's gonna be alright but they're gonna be annoyed by your lack of assistance
What do you choose?
74%White pill
26%Black pill
54 votes • Final results
Teague Lasser — e/acc retweeted
🚀🧠 Introducing Flow2: A Revolutionary Breakthrough in Neuroimaging!
Flow2 offers fMRI-like functional neuroimaging through TD-fNIRS + EEG integration. Portable & built to scale - let's unlock the potential of precision neuromedicine together!
#Neuroimaging #Biomarkers
Teague Lasser — e/acc retweeted
~gender || do you ever fantasize about the history of ideas (you are having an idea, someone is explaining an idea, pre-socratics, impact of ideas on behavior) during sex
37%M || fantasize abt ideas
32%M || no fantasy abt ideas
21%F || fantasize abt ideas
11%F || no fantasy abt ideas
38 votes • Final results Teague Lasser — e/acc retweeted
Releasing LLongMA-2 16k, a suite of Llama-2 models, trained at 16k context length using linear positional interpolation scaling. The model was trained in collaboration with @theemozilla of @NousResearch and @kaiokendev1. huggingface.co/conceptofmind…
Teague Lasser — e/acc retweeted
Announcing Nous Hermes Llama 2 13b!
This new version of Hermes, trained on Llama 2, has 4k context, and beats the benchmarks of original Hermes on every one we tested, including GPT4All benchmarks, BigBench, and AGIEval.
We have FP16, GGML, and GPTQ weights available!
Nous-Hermes-Llama2-13b weights have been released.
The model beats its previous benchmarks across the board, with the same dataset + hyperparameters.
4k context. GGML + GPTQ versions ready.
Thanks to our sponsor @RedmondAI for the generous compute.
huggingface.co/NousResearch/…
Teague Lasser — e/acc retweeted
Yoooo! Check out links.engi.network/cookbook to learn how to connect a wallet to Engi and use the CLI to create and complete bounties 🐉
This quoted post is unavailable.
Teague Lasser — e/acc retweeted
Happy #WhitePillWednesday
Among the likes of hunger, fear, reproduction, there is a deep and inseparable instinct. We yearn to create something we’ve never seen before. The desire to solve problems is innate to us, and it has taken us this far.
A glorious future it promises
Teague Lasser — e/acc retweeted
I've spent the last few months interviewing >60 experts in law, economics, AI, alignment, etc, on the impacts of AI, and safety interventions.
Today I'm publishing my first article, showing regulation designed to increase AI safety may backfire badly!
fast.ai/posts/2023-11-07-dis…
Teague Lasser — e/acc retweeted
All OpenAI's updates from today:
- Next week: Everyone gets code interpreter.
- From this moment: Everyone with a paying account gets API access to GPT-4
- And usage rate limit for paying accounts had been increased.
- Logprobs!! We finally get logprobs for the chat API endpoint!
- In the upcoming weeks - new model: gpt-3.5-turbo-instruct (Fine tuned version of gpt-3.5-turbo for instructions rather than chat.) (FINALLY!)
Later on - 2023:
- Fine-tuning API for GPT-4 & GPT-3.5.
2024 - Upcoming Deprecation:
- All of the first gen completion models will be deprecated on the beginning of 2024.
- Same goes for all the first gen embedding models.
- "Edit" API will also be deprecated.
[I really like it.. :(]
Teague Lasser — e/acc retweeted
Fifteen years ago this day, on 2012-07-01, the first hate crime was committed against a cyborg. It's #CyborgHateCrimeRemembranceDay.
Steve Mann was brutally and hatefully attacked in a McDonald's in Paris.
Attackers attempted to remove his surgically implanted EyeTap device.
ALT World's First Cyborg Steve Mann and his EyeTap Digital Eye Glass implant
ALT Anti-Cyborg Bigots