abdulkadir

@Rakkha72 @RnaudBertrand yeap, you are right... without that perfect English, they wrote perfect long paragraph with almost all English grammer rules applied.

@RnaudBertrand In Pakistan, people's English isn't always perfect even amongst the political class, it's actually very much possible that they do, or that it may have even been drafted by an AI.

This is just too funny 🤦‍♂️ Trump's justification for extending his deadline by 2 weeks is that Pakistan's Prime Minister "requested" it. But when you look at the edit history of the post in question by Pakistan's PM, where he "requests" the extension, it's painfully obvious it was sent to him by the White House since he first stupidly posted it with the mention "Draft - Pakistan's PM Message on X" 🤣

the new startup playbook looks NOTHING like the old one: – most of your team will be part-time contractors, creators, and ai agents – your first $1m will come from niching down. your next $10m will come from tastefully scaling out – one agent spins out 50 longtail SEO pages from transcripts, support tickets, or user reviews – startups are turning into QVC. except this time, you own the channel and the product – onboarding will feel like texting a friend. static forms are dead – every landing page rewrites itself based on who's viewing it (claude or chatgpt-4o + session data) – every successful company will feel like a subculture. the product is just a portal in – outbound are agents scraping, qualifying, and writing personalized intros 24/7 – customer support = 1 human backed by 5 lindy agents trained on every support ticket ever written – micro-apps will outperform mega-tools. specific > general – growth isn’t an afterthought. it’s built into the product (agent-invite loops, ai-powered referrals) – if your product doesn't spark curiosity in 2 seconds, it’s invisible – the best products of the next decade will be memes first, software second – “launch” is outdated. leak it instead – the new pricing model: $0 to play, $x to unlock identity – you won’t sell software. you’ll sell outcomes, transformations, identity upgrades – more people will leave big tech to build solo. not out of rebellion, but because their side hustles are more interesting – the best homepages become a scene. your standard shadcn websites won’t hit the same – default alive is low burn, small team, owned audience, high-leverage systems – competitor research happens automatically. agents scrape, cluster, and surface positioning gaps – your CRM isn’t stale. agents log calls, summarize deals, and write follow-ups before you hang up – venture capital is optional – customer success isn’t reactive. agents predict churn based on tone in support chats and usage – we’ll see more “tiny empires”: one founder, one audience, and a constellation of tools they own – bug reports are summarized, tagged, prioritized, and triaged by an agent before eng ever sees them – IRL matters. founders become event planners – most SaaS is overbuilt. the next wave wins by subtracting – if your product can't be explained in a screenshot, it won't spread – the creative director is the new power hire. taste is now a growth lever – churned users get a custom winback campaign built by an agent based on why they left – knowledge base builds itself from slack threads, loom links, and discord q&a (agents + gpt vision) – product feedback loops are instant. users speak → agents summarize, prioritize, and mock ui changes – most startups will die trying to be “all-in-one.” the winners do one weird thing stupidly well – startup advice used to be: find a technical cofounder. now it’s: find a distribution edge – your product isn’t finished when it works. it’s finished when people want to wear the hoodie – the people who win distribution will own demand. the rest will rent it if this felt like a glimpse into the future, it's because it is. instead of bookmarking this, share it with a friend, and start building. you don’t need permission to build like this. you just need to start. most people will ignore this. but this is the new reality... small teams, infinite leverage. Happy building. I'm rooting for you.

@AndrewYNg Necessity is the mother of invention. Every ambargo to China(or any other country) will make them better in that area. This is basic systems thinking. When changing policies, you have to make sure that long term implications will match what you seek for.

The buzz over DeepSeek this week crystallized, for many people, a few important trends that have been happening in plain sight: (i) China is catching up to the U.S. in generative AI, with implications for the AI supply chain. (ii) Open weight models are commoditizing the foundation-model layer, which creates opportunities for application builders. (iii) Scaling up isn’t the only path to AI progress. Despite the massive focus on and hype around processing power, algorithmic innovations are rapidly pushing down training costs. About a week ago, DeepSeek, a company based in China, released DeepSeek-R1, a remarkable model whose performance on benchmarks is comparable to OpenAI’s o1. Further, it was released as an open weight model with a permissive MIT license. At Davos last week, I got a lot of questions about it from non-technical business leaders. And on Monday, the stock market saw a “DeepSeek selloff”: The share prices of Nvidia and a number of other U.S. tech companies plunged. (As of the time of writing, some have recovered somewhat.) Here’s what I think DeepSeek has caused many people to realize: China is catching up to the U.S. in generative AI. When ChatGPT was launched in November 2022, the U.S. was significantly ahead of China in generative AI. Impressions change slowly, and so even recently I heard friends in both the U.S. and China say they thought China was behind. But in reality, this gap has rapidly eroded over the past two years. With models from China such as Qwen (which my teams have used for months), Kimi, InternVL, and DeepSeek, China had clearly been closing the gap, and in areas such as video generation there were already moments where China seemed to be in the lead. I’m thrilled that DeepSeek-R1 was released as an open weight model, with a technical report that shares many details. In contrast, a number of U.S. companies have pushed for regulation to stifle open source by hyping up hypothetical AI dangers such as human extinction. It is now clear that open source/open weight models are a key part of the AI supply chain: Many companies will use them. If the U.S. continues to stymie open source, China will come to dominate this part of the supply chain and many businesses will end up using models that reflect China’s values much more than America’s. Open weight models are commoditizing the foundation-model layer. As I wrote previously, LLM token prices have been falling rapidly, and open weights have contributed to this trend and given developers more choice. OpenAI’s o1 costs $60 per million output tokens; DeepSeek R1 costs $2.19. This nearly 30x difference brought the trend of falling prices to the attention of many people. The business of training foundation models and selling API access is tough. Many companies in this area are still looking for a path to recouping the massive cost of model training. Sequoia’s article “AI’s $600B Question” lays out the challenge well (but, to be clear, I think the foundation model companies are doing great work, and I hope they succeed). In contrast, building applications on top of foundation models presents many great business opportunities. Now that others have spent billions training such models, you can access these models for mere dollars to build customer service chatbots, email summarizers, AI doctors, legal document assistants, and much more. Scaling up isn’t the only path to AI progress. There’s been a lot of hype around scaling up models as a way to drive progress. To be fair, I was an early proponent of scaling up models. A number of companies raised billions of dollars by generating buzz around the narrative that, with more capital, they could (i) scale up and (ii) predictably drive improvements. Consequently, there has been a huge focus on scaling up, as opposed to a more nuanced view that gives due attention to the many different ways we can make progress. Driven in part by the U.S. AI chip embargo, the DeepSeek team had to innovate on many optimizations to run on less-capable H800 GPUs rather than H100s, leading ultimately to a model trained (omitting research costs) for under $6M of compute. It remains to be seen if this will actually reduce demand for compute. Sometimes making each unit of a good cheaper can result in more dollars in total going to buy that good. I think the demand for intelligence and compute has practically no ceiling over the long term, so I remain bullish that humanity will use more intelligence even as it gets cheaper. I saw many different interpretations of DeepSeek’s progress here in X, as if it was a Rorschach test that allowed many people to project their own meaning onto it. I think DeepSeek-R1 has geopolitical implications that are yet to be worked out. And it’s also great for AI application builders. My team has already been brainstorming ideas that are newly possible only because we have easy access to an open advanced reasoning model. This continues to be a great time to build! [Original text: deeplearning.ai/the-batch/issu… ]

@huybery @AndrewYNg I have been experimenting with different open models for classification of web interactions for some time. Qwen 2.5 was the best model so far even with a 7B parameters. Especially it is beyond the competition in prompt following. And now qwen 2.5 VL model... You rock guys 🙏

@AndrewYNg Thank you for recognizing Qwen!

Lololol

@elonmusk That’s why the fact you’re censoring righteous criticism of Israel shows to anyone who cares to look that you’re a liar of the highest order

Introducing Radio Hackernews! Turn articles into studio-quality podcasts in seconds. ◆ Fully autonomous, start to finish ◆ Highly reliable & completely serverless ◆ 100% open-source Built on Upstash Workflow. See how it works 👇🏻

When Palestine’s 🇵🇸 UN Ambassador Riyad Mansour asked the chilling questions that will echo through history to our eternal shame: “Is our blood cheap? Are our civilians less worthy of protection? Are our lives less sacred?”

My son Yahya, who was killed in by Israeli sniper near my destroyed house in May, and Israeli tanks ran over his body!

The Holodomor was a GENOCIDE committed by Jewish Bolsheviks. This madman was in charge of the genocide. His name is Genrikh Yagoda. He was Jewish. To those ignorant people claiming Holodomor wasn't perpetrated by Jews. They are 5% correct and 95% wrong. 95% of Bolshevik leaders were Jewish. 92% of the Gulag commanders were Jewish. 50% of the NKVD Generals were Jewish. The massmurder of at least 60 million white orthodox Christians and Muslims in USSR was committed by a 95% Jewish government. The Bolsheviks killed 60 million people and you never once heard about that in school!

"You can be super anti Muslim, you can talk about Muslims but as soon as you talk about Jews, you get cancelled. Why are they so protected?" - @DanBilzerian Bravo 👏🏾

“A minute of living with dignity and pride is better than a thousand years of a miserable life under the boots of the occupation.” - Yahya Sinwar

Piers Morgan has turned off the comments section on his Twitter after the Dan Bilzerian interview I wonder why?

@jakeshieldsajj Thank you brother, I’m honored by this award

I would like to congratulate @DanBilzerian on winning antisemite of the week When you win this award you know you are on the right side of history

This interview should be watched by the whole world. Piers Morgan

@Israel Fuck off nobody cares

Remember how nobody felt bad for Nazis? Well congrats Israelis, you’re the new Nazis And no amount of whining or pretending to be victims is going to make people feel sorry for you.

Not my best interview, but I said what needed to be said

@piersmorgan Call me what ever names you want, I say what I believe. You do what your told