imjonse

24 ms·

361.
by imjonse 3y ago
Of course it is working with local files, maybe I phrased it awkwardly. I am just not sure whether it gives you full context if you do that. I have used Cody a few times in the past few months and it's good, but at least its chat featu
362.
by imjonse 3y ago
This is written by a LLM pretrained on Steve's writings and then finetuned and made safe. A singe instance of "F-word" is the strongest it could muster.
363.
by imjonse 3y ago
Even though 'context' is mentioned many times in that post, I am still unsure if the power of Cody comes only if you import your repo in sourcegraph. It does not seem like using it within VSCode in your local source directory lets
364.
by imjonse 3y ago
BTW, Monsanto got rid of their infamous name by being bought and rebranded by Bayer.
365.
by imjonse 3y ago
Architecture details matter. "DeciLM-7B’s superior performance is rooted in its strategic implementation of variable Grouped Query Attention (GQA), a significant enhancement over traditional Multi-Query Attention (MQA) and standard GQA
366.
by imjonse 3y ago
If only they could raise a seed round and hire a decent team to add some animation on the frontpage and set it up on kubernetes then blog about their unique experience...
367.
by imjonse 3y ago
After ISL.
368.
by imjonse 3y ago
Jurafsky's 3rd edition draft of his NLP book and Simon Prince's DL book both have chapters on transformers, and the latter also on diffusion. Both have official free pdf versions. https://web.stanford.edu/~jurafsky
369.
by imjonse 3y ago
Murphy's book is about probabilistic/Bayesian machine learning whereas this, according to the preface is 'almost entirely non-Bayesian'. Murphy's book is also a lot more general, this one is about the deep learning
370.
by imjonse 3y ago
Tangential: one of Romania's main investigative journals, part of the Swiss Ringier group had just fired its main editors and around 20% of their staff last week, over their reluctance to stop investigating and writing about the local
371.
by imjonse 3y ago
Unless you read The art of war you cannot possibly become a 10x ninja developer. /s More seriously, I have read the design of everyday things about 10 years ago and it was one of the most boring books I have ever had to go through. I o
372.
by imjonse 3y ago
Hence the overwhelming number of hearts in the Twitter messages asking for the return of Sam Altman.
373.
by imjonse 3y ago
The source is gwern
374.
by imjonse 3y ago
Confusing name for the domain (Generally Intelligent) since it's the former name of a company in the AI/LLM area but does not seem to be related.
375.
by imjonse 3y ago
While this dataset is orders of magnitude smaller than what recent speech models like Whisper and Seamless got trained on, and while it is meant for supervised as opposed to self-supervised learning where data is more abundant, it can still
376.
by imjonse 3y ago
Indeed, Ctrl-F9 in Turbo Pascal was almost instantaneous, whereas the same in Turbo C++ would let you read through the list of files in the popup window as they were processed.
377.
by imjonse 3y ago
It also reads a bit like an official promotion for JAX (well-deserved btw) by mentioning its advantages separately and putting it first in the list of main frameworks.
378.
by imjonse 3y ago
"Keras 3 is a full rewrite of Keras" and "It's also a battle-tested framework" In case of a full rewrite how can we talk about battle-tested? I can understand the API is, but not the implementation.
379.
by imjonse 3y ago
"While we have demonstrated the efficacy of our simplifications across architectures, datasets, and tasks, the models we have considered (100-300M parameters) are small relative to the largest transformers." University researchers
380.
by imjonse 3y ago
I've run ollama on my macbook, watched a couple videos on prompt engineering, tried out stable diffusion on my phone. I am even working on a startup that is basically a shiny website plus an OpenAI API wrapper on the backend. What do y
381.
by imjonse 3y ago
He said he planned to a few months ago, but was busy/waiting to add some more features/fixes to the code first.
382.
by imjonse 3y ago
He probably has powerful connections beyond SV. He and Greg Brockman have been meeting heads of state and he has been fundraising in the Middle East recently. I wouldn't be surprised if he is sold as representing US interests, hence fe
383.
by imjonse 3y ago
Paul's tweet is an objective statement, it does not say anything about character or values and is not explicitly supportive.
384.
by imjonse 3y ago
Sam's integrity would be at home at Microsoft, for sure.
385.
by imjonse 3y ago
I don't obviously. But since those people were ready to jump ship to Microsoft, I am pretty sure they care more about their own careers than 'creating AGI that benefits humanity as a whole in the first place'
386.
by imjonse 3y ago
He may turn from powerful and well liked startup poster-child to simply powerful (like Larry Ellison, Bezos, Gates and countless other CEOs have in the past).
387.
by imjonse 3y ago
I can believe the staff likes or even loves him, but the following him part was mostly because of money/shares and because they know he's influential and well connected to people with money. And peer pressure may have had a part i
388.
by imjonse 3y ago
I wonder what is the rationale for picking a seasoned politician and economist (influenced deregulation of US finance system, was friends with Epstein, had a few controversies listed there). Has the government also entered the chat so obvio
389.
by imjonse 3y ago
Is being morally trustworthy a clear advantage from shareholders/profit pov? Those are what CEOs of companies are ultimately about regardless of mission/vision/PR of the company.
390.
by imjonse 3y ago
There are non-negligible chances that history will be written by Sam Altman and his GPT minions, so he'll probably be viewed favorably.

More ›