Shreyash

@samlambert 🤣

nobody technical buys this shit. only CIOs.

I passed the AWS Solutions architect associate exam last year I have some decent notes for the famous tricks that are used in the exam If some folks on my TL are interested, I can write an article about all those tricks (for revision) It won't be an elaborative one, just a last-minute helper for the exam

@gxjo_dev With political will, just 5 yrs But till then enjoy the show

How many years do you think it will take for India to produce a company on the scale of OpenAI or Anthropic?

Thread summary: • antirez: Chinese models aren't strong from distilling US ones via API. True distillation needs full logits (impossible at scale, esp. summarized CoT). • Reproducing complex model from few API points = mathematical sci-fi. • DeepSeek showed reproducible pretrain/SFT/RL pipelines; their distills helped thinking but weren't magic. • Black-box API use = SFT on outputs or RL signals (imitation), not proper KD. Limited transfer. • Replies: Anthropic warned on attacks; some call narrative cope, stress independent Chinese scaling & engineering. • TLDR: Stop the nonsense—it shows you don't get ML.

@antirez @grok Summarize the thread in one single tweet, use bullets

Another important thing: Chinese models are not strong because they distill US models. Distillation of models via API is *impossible*. If somebody tells you the contrary, they don't understand machine learning:

@sunnykgupta Indiranagar -> Rare, 45 min to 1 h max in worst case Kormangala -> Less frequent than BTM or HSR, but daily 30+ mins, worst case 2-3 hours 150AH would do for a 2 BHK

Folks who live in Koramangala or Indiranagar, how often does power outages last in these regions? Would a simple inverter work for most use-cases? Or is a generator required to have 24x7 power access?

@kingofknowwhere 🤣🤣

Unironically Bhavish's Krutrim was better than Sarvam at its best. If only it stuck to serving deepseek or kimi invoiced in rupees and re selling coding plans in rupees. The problem is that no one is doing this and unfortunately this is exactly what we need.

@elbhen14 @antirez @matteocollina India is a talent powerhouse, we just need educated policy makers I m hopeful that there will be top tier AI labs from India. @SarvamAI already shows a ray of hope

@antirez @WebDevCaptain @matteocollina Anywhere else than India would be a better option. The corruption there is off the charts and then the bureaucracy hurdles.

What Europe should do right now: 1. Call all the European researchers working on AI and return them back with same salary (or they can stay but switch career). 2. Fill EU places having GPUs with money, and put those people there. 3. AI partnerships with China + India.

@ergelgru @antirez @matteocollina You are wrong on "ability" It's just our Policy makers willingness to do theatrics than doing real work India has the talent and ability to create best in class models and applications

@antirez @WebDevCaptain @matteocollina A country that lacks the ability partnering with a continent that lacks the will It shouldn't be this way, it would be cool to live in a multi-polar world, but alas

@Hiteshdotcom Apne wale ko fark nhi padta

This is an alarm for every single country and coders. You might or might not have access to best tools in future

@mehulmpt Marketing blog post always takes time, it will sell Claude Code

Imagine telling somebody 5 years back that it now takes less time to rewrite a large codebase in another language than it takes to write a blog post about it

How can ppl judge someone without knowing them 😶 Its a fact on all social platforms that Girl's making these kind of videos will have more reach and engagement than boys, and everyone knows that But that doesnt mean that someone gets hired only bcz of this, it acted as a catalyst, but Skills >> Gender

Welcome to the team. She got hired because she brings so much to the table. The video got our attention, but her skills got her the offer letter. You have to appreciate the skill set; being dismissive of gender shows a pathetic mindset. > UI/UX > Graphic > Content creator > Confidence > Founder’s office experience > Worked around AI ads > Scriptwriter She had been applying for the last two months to every role that came up at TUF and reached out to so many at @takeUforward_ before the video even. Hence, it was a no-brainer for us when she applied for the Growth role. Welcome, let’s build @takeUforward_ together and take it forward 😇

@Hiteshdotcom 200$ is the new normal if u wanna use it for real tasks

Claude Fable 5 is really nice but barely useable in 100$ plan. I am constantly hitting limit. How is your experience with Fable?

Imagine, College is focused just on studies, research and building great things in every branch. Campus placements are not allowed, they cannot invite companies neither can advertise about placements, just pure education, including IITs n private colleges. Thoughts on this fictitious situation 🤔

More AI-generated code doesn't make your team faster. It might actually slow you down.

I m always betting on Gemini in this LLM/LVM (Foundation models) race but after Fable 5 (Mythos) release, I m not afraid to share this meme, IYKYK #Fable #Mythos #Claude #Gemini

@lydiahallie And its slow, very slow

Claude Fable 5 is now available in Claude Code! 🎉

@turiyatman Yeah, agreed

Tomorrow I get my salary hike. I already know the exact number. 😊 No appraisal system. No ratings. I work in a Norwegian organization where everyone gets the same % annual raise. When I first heard that, I wondered: Then why would anyone work hard? What surprised me was the opposite. People work hard anyway. When you're not chasing ratings and competing with colleagues, you stop grinding for the appraisal and start working for the actual goal. Your thoughts, is this okay?

@claudeai I discovered this on Claude console and not on twitter 😂

Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use. Its capabilities exceed those of any model we’ve ever made generally available.