I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
I think the most valuable experience you can gain right now as a software engineer is to build a project in a language you donโt know.
Constraint breeds creativity. You will have to learn new ways of working with AI when you realise you canโt write or review code by hand.
How will you package work? Set goals? Know where and when to refactor? How will you find bugs and validate fixes? How will you gain and keep understanding over your codebase?
Anthropic have started as the good boys of AI.
But seeing their stance on pirated content made me think that they are the most evil one.
According to them they can take any data available on internet or on earth, like its their right.
But when someone creates an open source model by distilling their output, they start lobbying US government and want these open source model to be banned in US.
@deedydas no entrepreneurs and sportsman in this?? because entrepreneurship and sports person are a new age profession (major boom in last 20-30 years)??
This line from Atomic habits don't leave my brain.
It doesn't make sense to continue wanting something if you are not willing to do to what it takes to get it. If you don't want to live the lifestyle then release yourself from the desire
BREAKING: nearly every major US AI company just signed a letter from NVIDIA asking Washington not to restrict open-weight models:
Jensen Huang posted it on July 24, his first post ever on X. It hit 11 million views and pulled 50 signatures within a day. It's now past 75.
The letter asks Washington not to restrict models anyone can download, inspect, modify, and run on their own infrastructure. The trigger was the White House floating action against Moonshot AI, maker of Kimi K3, the strongest open-weight model in the world.
Signers include NVIDIA, Google, OpenAI, Meta, Microsoft, AMD, GitHub, Palantir, Hugging Face, a16z, and Y Combinator.
Why sign: these companies make money when AI spreads widely and cheaply. Chipmakers sell more chips. Clouds rent more compute. VCs want startups building on free models instead of paying frontier prices.
Restrictions would cut into all of it. Framing it as American competitiveness is how you get Washington to leave it alone.
Two big names are still missing: Amazon and Anthropic.
Anthropic sells closed models and has warned for years that released weights can't be pulled back. Amazon is Anthropic's largest investor, and Anthropic runs its models on Amazon's Trainium chips. Google, also an Anthropic backer, signed anyway.
What it means if you build on AI: this fight decides what you can run and what you'll pay for it. Restrictions push everyone toward a few closed providers and their pricing. That's the real stakes under the "American competitiveness" framing.
Your thoughts?
is AI quietly becoming the first person we ask for anything now?
Incident - 1
Yesterday, I went to a cafe to work.
The person sitting next to me was preparing for a software interview. He had ChatGPT open and was asking it to explain a system design question.
Incident - 2
A few minutes later, someone behind me joined a work call.
His teammate was explaining a customer problem. He listened for two minutes and said: "Put the feedback into Claude and ask it to find the common complaints."
Incident - 3
Around lunchtime, I met a friend.
He told me he had started following a new workout and diet plan. I asked him who created it. He opened Gemini and showed me the conversation.
Incident - 4
In the evening, I went to play FIFA with friends.
One of my friends wanted to text the girl he is seeing about changing the timing of their next date. He wrote two lines, then opened ChatGPT and asked it to make the message sound more polite.
whats your take on this??
I was in the office and one person said,
"AI will replace humans"
Another person replied : Not possible, the most important human skill is listening
so I used claude to learn this more and then claude gave me 1000 words output.
And then claude asked me - Would you like me to expand further.
#humor #AIvsHumans
This @ycombinator startup may have just changed its strategy to grow against its rival @WisprFlow .
Willow Voice made unlimited AI dictation free forever.
That looks like a pricing change.
I think it is a bet on the future of voice.
AI dictation is getting cheaper, faster and increasingly commoditized. Soon, charging users simply to convert speech into text may become difficult.
So Willow appears to be moving the paywall:
โ Give away basic dictation โ Build a daily voice habit โ Grow faster through free usage โ Charge for accuracy, personalization and advanced workflows
This fits an idea shared by @gokulr :
Not every product should be monetized. Some products should exist to drive retention.
Willow is effectively saying:
Donโt monetize voice input. Own the voice input habit.
That puts pressure on Wispr Flow, which still limits free usage and charges for unlimited dictation.
It may be the company that becomes the default keyboard before voice dictation becomes a commodity.
Is Willow giving away too much or has Wispr Flow placed its paywall on the wrong layer?
Learning from youtube, Coursera, Udemy but still not able to understand concept and built and Intuition around it.
Try skylab real time tutor who talks and draws while teaching to you .
https://t.co/H8x2LsbBk4
Learning from youtube, Coursera, Udemy but still not able to understand concept and built and Intuition around it.
Try skylab real time tutor who talks and draws while teaching to you .
https://t.co/H8x2LsbBk4