30years ago, murder of Tamil school girl Krishanthy Kumaraswamy led to discovery of Chemmani mass graves. 600 human remains have been uncovered, making Chemmani largest mass grave in Sri Lanka. Tamil victims and families of disappeared deserve truth, justice & accountability. 1/2
Chinese researchers published a paper with a devastating title: "The End of Software Engineering”
it argues software engineering is finished.
In traditional software, code is the carrier of pre-written human logic.
In agentic software, the AI agent is the software.
Code is no longer a permanent monument built by human hands. It is completely ephemeral, dynamically generated, executed, and discarded on the fly by an LLM-driven reasoning loop.
Think about how software delivery has evolved:
• Era 1: On-premise licensed software (you installed it locally)
• Era 2: SaaS (hosted in the cloud, managed by vendors)
• Era 3: Agent-as-a-Service (AaaS)
Each historical shift transferred complexity away from the user. But this latest shift transfers something entirely different.
It transfers decision-making complexity itself.
The paper argues that traditional engineering is hitting a hard complexity wall. Human brains can only hold so much state, manage so many dependencies, and debug so many lines at once.
LLM-based agents scale non-linearly.
They don't just write functions faster. They navigate architectural complexity by outsourcing reasoning to models that improve every single month.
Which means the role of the developer is permanently changing.
You are no longer a code author typing syntax line by line.
You are an intent architect.
Your job is no longer writing the implementation. It is specifying goals, designing multi-agent coordination loops, and auditing outcomes.
The remains of 582 people, including babies and young children, have been uncovered from a mass grave in northern Sri Lanka.
The site is linked to allegations that the country’s military killed Tamil civilians during the Sri Lankan civil war, which lasted from 1983 to 2009 ⤵️
https://t.co/OgxdlwsBLe
Anthropic just told the world to slow down AI development.
The same day they disclosed Claude wrote 80% of their own codebase.
The same week they filed for a $1 trillion IPO.
The same month they gave a self-escaping AI to the EU’s cybersecurity agency.
Their engineers now ship 8x more code per quarter than they did last year.
The AI is building the AI that builds the AI.
Anthropic’s solution is a globally coordinated pause.
Requiring China to agree.
Voluntarily.
The company that cannot stop building faster is asking everyone else to stop building faster.
BREAKING: Anthropic has urged for a global pause in AI development as artificial-intelligence models are nearing capability to improve without human intervention, per WSJ
#chip_wars#AI_stocks#custom_asics
AI race-ல NVIDIA மட்டும் தான் ராஜா மாதிரி தெரியுது. ஆனா உண்மையில hyperscalers (Google, Meta, Amazon) ஒரு silent shift பண்ணிட்டு இருக்காங்க
GPU monopoly-யை உடைக்குற game தொடங்கிடுச்சு!
இதுவரை NVIDIA GPUs எப்படி AI infra ல core compute layer ஆக மாறிப்போனதுனு
பேசிருக்கோம்.
👇
https://t.co/fqzEc7y4bk…
hyper-scalersக்கு (Google, meta.) , தங்களோட லாபத்துல ஒரு பகுதி NVIDIA கிட்ட போய் சேருது. அது இல்லாம Nvidia மட்டுமே dependency ஐ உடைச்சு ஆகணும்.
இதனால , market la custom ASICS (Broadcom ) பிரபலமாக ஆரம்பிக்குது.
GPU க்கும் , custom ASICS க்கும் வித்தியாசம் தெரிஞ்சிக்கனும் ..
GPU: ஓரே வேலையை , பல பேர் , ஓரே நேரத்தில செய்ய வைக்கும்.. வேலை என்ன வேணாலும் இருக்கலாம் .
AI model training , நாம போன thread la பேசுன மாதிரி , run the model against training sets using different weights in parallel.
ASICS: ஒரு குறிப்பட்ட வேலைய திறன் பட செய்யும்.எல்லா வேலையும் அல்ல .
Inference engine - நாம daily chat gpt, Gemini கிட்ட கேட்கிற requests , இதான் process பண்ணி பதில் தருது.
Google Image rmodel , இதோட வேலை image recognize பண்றது மட்டுமே. நாம கொடுக்கிற image ஐ கண்டுபிடிக்கிறதுல specialist .
இப்போ training mostly GPU, but hyperscalers slowly ASIC trainingக்கும் move ஆகிறாங்க..
இப்ப,
Google மாதிரி hyper scalers க்கு LLM model training க்கு Nvidia உம், Inference engine (நாம கேட்கிற கேள்விகளுக்கு பதில் சொல்ற LLM process ) க்கு custom ASICs உம்
அதிகமா உபயோகப்படுது .
Custom ASICs 30-70% compute investment ஐ குறைக்க உதவுது.
அதனால தான் Amazon (inferentia chips) , Google (TPUs) .. hyper-scalers custom ASICs நோக்கி படையெடுக்கின்றன.
Broadcom- Hyper-scalars க்கு custom silicon design partner.
Google / Meta மாதிரி players → in-house chip, but designஐ outsource பண்ணிடுவாங்க
Broadcom மாதிரி companies கிட்ட.
Broadcom stock went from 800$ to 4200$..
100% YOY growth …Marwell, AMD … already running parabolic 📈
சரி அப்ப யார் தான் இந்த race ல ஜெயிக்கிறா?
1. ASICs , Nvidia replace பண்ணாது.
2. NVidia growth slow down ஆகும்.
3. Mix of ASICs and Nvidia GPUs in the ground in future .
Chips வேணா Nvidia, Broadcom ன்னு
Players இருக்கலாம், ஆனா factory நடத்துறது
இந்த race ஓட ராஜா நம்ம , Hyper-Scalers தான் !
தொடர்ந்து பேசுவோம்..
Announcing Amazon S3 Files.
The first and only cloud object store with fully-featured, high-performance file system access.
Learn more here. https://t.co/rNuWa5Rsi2