Michael Viewer

Empowering homeschoolers with AI 🚀 Creator of a multi-agent AI tutoring system covering all subjects & aligned with the core curriculum.

[Michael Viewer @Michael46813795]

I’m a linguist, teacher & EdTech builder. Since 2021, I’ve homeschooled my son, testing endless methods & tutors to cover the core curriculum. Nothing was perfect.
In 2025, AI agents finally reached the level I needed. So, I built the ultimate solution. 🧵👇

[Michael Viewer @Michael46813795]

@turningpointjpn On tutaj wydyma powietrze w policzkach a nie trzyma tam jedzenie :)

[Michael Viewer @Michael46813795]

@FinansowyUmysl Ważniejsze jest to, jak się uczyć.

[Michael Viewer @Michael46813795]

@HamdanMohammed Big companies will come anyway.
The real opportunity is attracting solo AI builders and small teams early.
With the right support (grants, competitions, low-barrier setup), UAE could become the default place to start in AI. Good luck! 🙏

[Michael Viewer @Michael46813795]

@lovindubai Shorter school days sound great, but kids fall out of routine fast.
I’ve been keeping it simple with short lessons at home using Tutor Michael Ai, works great for me.

[Michael Viewer @Michael46813795]

@gulf_news @lewy_official is the best, great career.

[Michael Viewer @Michael46813795]

@aalbeirutty Dopiero użycie systemu wieloagentowego pozwala osiągnąć poprawne rezultaty.

[Michael Viewer @Michael46813795]

@mordownik4u Żarty z pogrzebu sobie robią.

[Michael Viewer @Michael46813795]

@mindsetmachine This is exactly why I have developed https://t.co/An8PqJUyL9, tech is not a problem.

[Michael Viewer @Michael46813795]

It solves the biggest homeschooling headaches:
📊 Precise progress tracking & weakness detection
📝 Detailed "parent-teacher" reports (no slacking off!)
🎯 100% aligned with official standards

Get your time back while they learn. Check it out: https://t.co/uJLfHG6wsu 🚀

[Michael Viewer @Michael46813795]

I’m a linguist, teacher & EdTech builder. Since 2021, I’ve homeschooled my son, testing endless methods & tutors to cover the core curriculum. Nothing was perfect.
In 2025, AI agents finally reached the level I needed. So, I built the ultimate solution. 🧵👇

[Michael Viewer @Michael46813795]

Meet https://t.co/uJLfHG6wsu 🤖 – a multi-agent AI education system.
Unlike passive tutoring, it uses the Socratic method to force active thinking. A digital team of AI agents guides the student step-by-step through the official core curriculum, adapting to their pace.

[Jakub Mrugalski 🔥 @uwteam]

Mówi się, że modele językowe (LLM) dają lepsze wyniki, gdy pozwolimy im "myśleć na głos". Wiemy, że LLM-y słabo radzą sobie z matematyką.

Postanowiłem więc zrobić prosty test...
Matematyczny oczywiście 😈 🧵 ↓

[Michael Viewer @Michael46813795]

@Polsport W takich okolicznościach to wyłącznie prawo Celsjusza.

[Robert Bortins @TheRobertBshow]

Your child's education is being scheduled by a bureaucrat who has never met them.
Meanwhile, homeschool families are starting their day when their children are actually ready to learn.

[Ethan Mollick @emollick]

Nano banana turns out to be Gemini Flash 2.5 Image Generation (not quite as catchy a name). I had a bit of early access, first through the same LMArena link everyone had, then privately.

It is impressive, crossing a threshold that goes beyond toy (though is a pretty fun toy too)

[Eric Wallace @Eric_Wallace_]

Today we release gpt-oss-120b and gpt-oss-20b—two open-weight LLMs that deliver strong performance and agentic tool use.

Before release, we ran a first of its kind safety analysis where we fine-tuned the models to intentionally maximize their bio and cyber capabilities 🧵

[Alex Albert @alexalbert__]

Say hello to Claude Opus 4.1, our best model yet for real-world coding, reasoning, and agentic tasks.

The frontier continues to move forward one update at a time.

[Psyho @FakePsyho]

Humanity has prevailed (for now!)

I'm completely exhausted. I figured, I had 10h of sleep in the last 3 days and I'm barely alive.

I'll post more about the contest when I get some rest.

(To be clear, those are provisional results, but my lead should be big enough)

[Simon Willison @simonw]

I think context engineering is going to stick - unlike "prompt engineering" it has an inferred definition that's much closer to the intended meaning, which is to carefully and skillfully construct the right context to get great results from LLMs

[Rohan Paul @rohanpaul_ai]

These guys literally burned the transformer architecture into their silicon. 🤯

And built the fastest chip of the world of all time for transformers architecture.

500,000 tokens per second with Llama 70B throughput. 🤯

World’s first specialized chip (ASIC) for transformers: Sohu

One 8xSohu server replaces 160 H100 GPUs.

And raised $120mn to build it.

🚀 The Big Bet

@Etched froze the transformer recipe into silicon.

By burning the transformer architecture into its chip means it can’t run many traditional AI models: like CNNs, RNNs, or LSTMs. also it can not run the DLRMs powering Instagram ads, protein-folding models like AlphaFold 2, or older image models like Stable Diffusion 2.

But for transformers, Sohu lets you build products impossible on GPUs.

HOW ❓❓

Because Sohu can only run one algorithm, the vast majority of control flow logic can be removed, allowing it to have many more math blocks.

As a result, Sohu boasts over 90% FLOPS utilization (compared to ~30% on a GPU7 with TRT-LLM).