Sandesh

Sr. Scientist @amazon | Computer Science graduate from @OhioState | Linguistics, ML and #NLProc | Social Media analysis... | Opinions my own, not my employer's

[Sandesh @padfoot_29]

@rasbt Ordered it off Manning and waiting for the physical copy. Congratulations! 🙌

[Tim Dettmers @Tim_Dettmers]

My new blog post discusses the physical reality of computation and why this means we will not see AGI or any meaningful superintelligence: https://t.co/jsAKQ6T3gC

[Sandesh @padfoot_29]

@AmandaAskell Never thought that Bernie Sanders of all people will have this take. Shocking.

[Sandesh @padfoot_29]

@MichaelVaughan Clearly, clickbait headline. There was no "bizzare tirade". Asking for consistency in the use of technology is not bizzare. Neither did he say it wasn't out?

[Sandesh @padfoot_29]

@thamar_solorio @emnlpmeeting @mbzuai That's fantastic! 😮

[Omar Khattab @lateinteraction]

Everyone talks about AI engineers but no one appreciates the poor AI plumbers fighting leaky abstractions all day long

[Jaana Dogan ヤナ ドガン @rakyll]

Burnout happens when there is no progress. Burnout also happens when you have no visibility into how dots will connect in the future. Burnout is about not being able to close the loop after putting a ton of effort and not seeing enough results.

[Rosanne Liu @savvyRL]

What the hell is "agent"? Is it just a software system where you replace lots of code with "let's ask an LLM"?

[Christopher Manning @chrmanning]

There are 2 mistakes you can make about LLMs:

① Thinking everything LLMs say is correct, they can reason, and with a bit more scale they’ll get us to superintelligence

② Thinking LLMs are good for almost nothing—they are FAR better at all #NLProc tasks than previous methods

[rohit @krishnanrohit]

Except for pollution, India has leapfrogged the West in almost everything, at least within the upper middle class

[Kevin Patrick Murphy @sirbayes]

While GenAI is fun, I think its economic value is grossly over estimated, because it’s unreliable, risky and expensive to make and serve. It’s fine for creative tasks, but not (yet) autonomous agents

[Sandesh @padfoot_29]

@narges_tabari @abacusai Indeed! Thank you :)

[Andrej Karpathy @karpathy]

In AI at least, the real 30 under 30 imo you have never heard of. They are 5 layers down the org chart from the CEO. They are usually not on Twitter, they have an unmaintained LinkedIn, they don’t go on podcasts, and they maybe published at one point but don’t do so anymore. They are today’s Einsteins and directly invent and build miracles. I wish they weren’t hidden inside big tech and startups in return for very large paychecks but here we are.

[Sandesh @padfoot_29]

Today's sunset from Golden Gardens park. It only kept getting better and this picture doesn't do it justice. #pnw #sunset #seattle #winter #goldengardens #pnwonderland

[Graham Neubig @gneubig]

ACL has removed the anonymity period.

This means that ACL submissions can be posted and discussed online at any time, although extensive PR is discouraged.

https://t.co/F7xVdCEoyY

[Sandesh @padfoot_29]

This is actually really cool! ~3B model, but on the few benchmarks here, it pretty much beats every other model (and is close to llama-2-70B).

[elvis @omarsar0]

Don't sleep on small models! Microsoft clearly isn't.

[Sigma Sreedharan @sigmas]

Good Morning, #Seattle!!!! Now that’s what I call a #sunrise!!! #wawx. An iPhone video for now. Click to view full length. #wawx

[Narges Tabari @narges_tabari]

This is a joint work with my colleagues at AWS AI Labs - @padfoot_29 , and Rashmi Gangadharaiah.

Link to the paper: https://t.co/WjgdsOda9b.

#nlproc #recommendersystems #languagemodels #personalization 5/5

[clem 🤗 @ClementDelangue]

My prediction: in 2024, most companies will realize that smaller, cheaper, more specialized models make more sense for 99% of AI use-cases. The current market & usage is fooled by companies sponsoring the cost of training and running big models (especially with cloud incentives).

[Sandesh @padfoot_29]

@GuggerSylvain All the very best for what comes next, Sylvain! I learnt a lot from your initial notes and lectures on the HF website when things were just picking up.