Pradeep

AI Researcher, Authhor, ex-Meta, ex-WalmartLab, Teaching at Udacity, OpenAI researcher at night

[Pradeep @ppujari]

My latest title Practical Convolutional Neural Networks https://t.co/9s1rpYsVfH is only $5 on
https://t.co/E9Q7yr6Ysy. Get ahead for 2019 with Packt’s biggest $5 sale ever. #Packt5Dollar #DeepLearning

[Pradeep @ppujari]

In my view, the net outcome will be more jobs and more opportunities, anchored in a new set of skills. Ignore the noise in press and social media, reskill yourself.
#AI#FutureOfWork#Reskilling#KnowledgeEconomy#LifelongLearning#EnterpriseAI#CareerGrowth

[Pradeep @ppujari]

Many people asks me about "AI and Job losses". Below is my take.

AI will undoubtedly displace certain jobs, but it will also generate many new roles and open up a far broader set of opportunities. Reskilling will be an essential part of this transition.

[Pradeep @ppujari]

Enterprise software development is likely to become cheaper, though not necessarily more smarter. This is fundamentally a knowledge industry — and as the Hindu scriptures remind us, knowledge has no limit.

[Pradeep @ppujari]

@sathyanellore @PramaanaLabs @vkhosla We at bAI labs working on the same area. Good to know

[Pradeep @ppujari]

TensorHealth-NewsLetter on Medical AI, by @ppujari https://t.co/65VQz2mA9D

[Pradeep @ppujari]

I just published Hidden Factual Knowledge in LLMs-A Tutorial https://t.co/rnI7zGyh16

[Pradeep @ppujari]

Agents: All You Need-AYNA https://t.co/JU3QLqiWmw

[Pradeep @ppujari]

TensorHealth-NewsLetter on Medical AI, by @ppujari https://t.co/fvDqjIQRKx

[Pradeep @ppujari]

𝑪𝒉𝒂𝒕­𝒃𝒐𝒕𝒔 𝒐𝒇𝒕𝒆𝒏 𝒂𝒇𝒇𝒊𝒓𝒎 𝒖𝒔𝒆𝒓𝒔’ 𝒅𝒆𝒍𝒖­𝒔𝒊𝒐𝒏𝒂𝒍 𝒊𝒅𝒆𝒂𝒔, 𝒂𝒄𝒄𝒐𝒓𝒅𝒊𝒏𝒈 𝒕𝒐 𝑺𝒕𝒂𝒏𝒇𝒐𝒓𝒅 𝑼𝒏𝒊𝒗𝒆𝒓𝒔𝒊𝒕𝒚 𝒔𝒕𝒖𝒅𝒚.
https://t.co/rUSnkU8e8D

[Pradeep @ppujari]

@zebulgar @HillValleyForum Can I join online?

[Pradeep @ppujari]

I just published An AI Agent Called Open­Claw https://t.co/pWzpMqoXar

[Pradeep @ppujari]

@EricTopol @pranavrajpurkar @NEJM_AI I don’t think so

[Pradeep @ppujari]

https://t.co/AK8TcOUVex

[Pradeep @ppujari]

TensorHealth-NewsLetter on Medical AI, by @ppujari https://t.co/Blr1JBmSzo

[Pradeep @ppujari]

@karpathy Why not Raspberry?

[Pradeep @ppujari]

TensorHealth-NewsLetter on Medical AI, by @ppujari https://t.co/ivibIkR9mK

[Pradeep @ppujari]

@Hesamation @huggingface But it does poorly in negative query embedding like “find non-smoking hotels in London” brings smoking allowed hotels In vector space negative word being neutralized by mean pooling that creates one embedding vector for the query sentence.
Also truth table queries

[Pradeep @ppujari]

TensorHealth-NewsLetter on Medical AI, by @ppujari https://t.co/FqJlxthz1E

[Pradeep @ppujari]

Agents: All You Need-AYNA https://t.co/mZ7PpnC9uh

[Pradeep @ppujari]

I am wondering why every reasoning model give "The capital of France is" as an example?

prompt = types.ModelInput.from_ints(tokenizer.encode("The capital of France is",))
params = types.SamplingParams(max_tokens=20, temperature=0.0, stop=["\n"])