Tang U-Liang

Founder Bezalel Consulting | EduTech Solo Builder & Consultant building educational SaaS to help Malaysian and SEA educators save time and boost outcomes

[Tang U-Liang @uliang6482]

So you're saying that Anthropic is trying to recall Fable's rollout, and using the US government as a convenient scapegoat..."See they made us do it!" .

Plausible, and so on brand for the big AI labs.

[Xenova @xenovacom]

I gave Fable 5 one job: write custom WebGPU kernels for Gemma 4 inference.

It climbed to 84 tok/s, then hit a wall, insisting further optimization was impossible.

Hours later, Anthropic rolled back invisible LLM development safeguards, and it hit 255 tok/s.

The next day, access to Fable 5 was suspended globally.

[Anthropic @AnthropicAI]

The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.

The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.

Access to all other Claude models is not affected.

We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.

Read our full statement: https://t.co/bwn0sximKZ

[Tang U-Liang @uliang6482]

IT and software engineering in general is a thankless job. If the light's keep on blinking, nobody's gonna notice.

But everybody notices if it does.

[Dangerous Thoughts @DangerousThinkg]

The story of my professional life told in 54 seconds

No matter how good the pre-planning, no matter how organized it starts out, this is where I end up

[Tang U-Liang @uliang6482]

I think I bumped into his recently at Kino KLCC. He was talking to two young (students?). His security officer was helping him with some books. I was wondering why his face looked so familiar, until I realised that his face is displayed over every hospital and government office in Ipoh.

Very understated man. It was surreal, and a realization how different things are from the times of the Malacca Sultanate, where Sultan Alauddin Riayat Shah would need to disguise himself as a commoner to be around the rakyat. Now, he has the freedom to browse in a bookstore and people won't bother him.

[Tang U-Liang @uliang6482]

https://t.co/K9GHoFFtLS

[Ihtesham Ali @ihtesham2005]

David Epstein studied the world's best athletes, scientists, and inventors, and found they all broke the same rule.

Here are 10 reasons from "Range" why generalists beat specialists in everything that matters.

1) Specializing late is an advantage, not a delay

[Tang U-Liang @uliang6482]

Intelligence is collaborative. Different fields cross fertilize. Intelligent people recognize common patterns from different areas of life and apply them to another. This has been the pattern of invention and discovery. Synthetic intelligence is no difference.

See link below

[Nicolas Bustamante @nicbstme]

What I find fascinating with Claude Fable 5 is it proves once again that large generalist models will outperform vertical ones.

On ProofBench (graduate-level formal math benchmark in Lean, where a proof either compiles or it doesn't) Fable 5 beat Harmonic's Aristotle, 77% vs 71%.

Aristotle is a system built specifically for formal math + run on its own internal harness, so the generalist beat the specialist on the specialist's home turf.

It's the Richard Sutton's "The Bitter Lesson". His whole argument is that across 70 years of machine intelligence research, the methods that win are the general ones that scale with compute. Not the ones where we hand-encode human expertise. Building our own knowledge into the system feels good and helps short term gains but long term it always gets overtaken by bigger model.

You can look at Chess, Go, speech, vision, same story every time. First the specialized model wins, then the general one takes over.

and btw this is the whole premise of AGI. You don't build one model for math, one for code, one for law. you build a single general model that scales with compute and it learns to do everything

[Tang U-Liang @uliang6482]

It's like everyone is rediscovering the basics of programming once again, as agents predictably become an abstraction over structured programming.

What's next? Pipes?

[Matt Pocock @mattpocockuk]

Everyone's banging on about loops

When they should be thinking about queues

[Tang U-Liang @uliang6482]

The entire edifice of the Internet economy is run on the back of Sir Tim Berners-Lee who collects no royalties for the utilisation of the http protocol which he invented.

The value of an invention does not necessarily accrued to its creator, but that does not necessarily mean the effort wasn't worth it.

[Gabriel @gbrl_dick]

if this guy was born in brookline, mas he’d be worth 35 billion dollars and an entire field of science would be funded solely by his personal foundation

[Tang U-Liang @uliang6482]

This is directionally correct and I'm tweaking my workflow to have one session where we convert specs into tests. And then start a loop or /goal to implement until tests pass. The implementor cannot alter the tests.

[Kun Chen @kunchenguid]

oh gosh, this is kind of a big deal

DO NOT ASK YOUR AGENTS TO DO TDD!

i now have empirical evidence that Test Driven Development is harmful for coding agents

what other popular skills do you want me to debunk?

details about the evaluation below 👇

[Tang U-Liang @uliang6482]

I have been saying this for a while, but DAP, and PH in general cannot rely on the PAS = taliban fear to take the non Muslim vote for granted. Sooner or later, voters will realise that the best way to punish the governing party is to vote the previously unthinkable.

[Noor Suhana Ishak 🇲🇾 🏴 @cek_suhana]

My supervisor, a non muslim & was Teresa Kok’s voter, said to me that for the next election if PAS is going to contest Seputeh, she will vote. Then I told her abt the non muslim wing in PAS. She was so excited and asked where she can register.

This might seems like a joke, but

[Tang U-Liang @uliang6482]

AI video generation will be the next frontier of AI assisted productivity.

As you see from his explanation. Prompting is neither random nor a pulling on a slot machine, but amplifies the directorial and artistic voice.

Just like coding, you have to know what you are doing (not that I am any good at video creation) in order to get good results.

It's only a matter of time before AI gen storytelling will be the breakout media format.

[AYi @AYi_AInotes]

全网爆火的《丧尸清道夫》作者刘梓渝公开了自己的视频创作思路，内容非常干，值得所有AI短视频创作者逐帧学习！

先说结论，跟抽卡运气没太大关系，
核心是把自己的导演思维、镜头语言和后期剪辑，
完整嵌进了AI工作流。

具体视频内容和我做的总结：⬇️

[Tang U-Liang @uliang6482]

@PaulGugAI Depite people panning Apple for not getting into the model race, I think they have made the smartest AI play. The hardware.

[Tang U-Liang @uliang6482]

@PaulGugAI The rest of the channel is a treasure trove of memes for the AI age

[Tang U-Liang @uliang6482]

How it feels like when the conversation is now turning to running agents on loops.
https://t.co/v6YsewYyx3

[PJ Ace @PJaccetturo]

Ladies and gentlemen, it’s here:

I’m proud to announce that 'Nexus' will be my upcoming hybrid feature film.

Here is a 5-minute teaser, made by 3 people in 2 weeks.

Made with Dreamina AI using Octo & Dreamina Seedance 2.0, full workflow coming soon

[Tang U-Liang @uliang6482]

@AriSchulman I have an autistic non verbal son. It is clear to me and many other parents in the same boat that thought and speech are two very different things. One can exist without the other.

[Tang U-Liang @uliang6482]

Since we're swapping VC stories, I have my own experience trying to raise funds at pre-seed level for Temansihat. Not VC, but from my cofounders circle of contacts.

At the time, we were working on the assumption that a medical assistance use case by a chatgpt like interface would be compelling. I came up with the concept that chatGPT could drive other side effects on a user's phone, notifications, trigger widgets besides just chat etc by returning json structure to the client aka. tool calling but on the front end.

That was last year, and I was much less knowledgable then about AI and the advancements since. I still hand coded many things, and only now did I realize that what I was trying to reinvent was an agent harness. And there were already frameworks that (i.e. LiveKit) implemented my concept.

Anyway, we didn't succeed in raising the targeted 1.2 mil, and most of the potential investors couldn't understand the POC I put together. Some remarked after wards, "You had better hire a better UI designer." (Which we did eventually, but he still wasn't very good).

But it was a great learning experience and that really opened my eyes to the level of expectation people have in a product before investing.

[GREG ISENBERG @gregisenberg]

I was once pitching in a board room at a top 3 VC firm for a $15M Series A.

12 people in the meeting. One of the GPs fully fell asleep. Out cold for 30+ minutes. Nobody acknowledged it. Everyone just kept going.

I kept presenting my Series A slides to an unconscious man in a Herman Miller chair and somehow that was considered normal. That's venture capital.

You might fly across the country to perform for people who may or may not be conscious.

It's a dance.

And sometimes you lead and sometimes you follow and sometimes your partner is unconscious.

If you're raising right now, just know: every founder has a story like this. The process is weird. The power dynamic is weird. You're not crazy for thinking it's weird.

No one talks about it because they want to continue raising. But I'm happy to stick my neck out there.

It is weird.

[Tang U-Liang @uliang6482]

Every founder has their weaknesses.

Steven Jobs was known to be emotionally abusive, Elon is autistic, Dario has a God complex, Sam A. is an allegedly shifty snake oil salesman...

I try to overlook all these. I use a macbook, hoping to get a Tesla soon, subscribe to Claude and thing GPT is a revolutionary piece of tech.

Wish I could try Codex, but can't afford two subs 😅

[Tang U-Liang @uliang6482]

@PaulGugAI definately!

[Tang U-Liang @uliang6482]

I'm going to say that it's rare to come by leaders (in Malaysia) who take this approach to conflict resolution.

The culture here is pretty much "I'm the boss, I've eaten more salt than you've eaten rice, so it's my way or the highway." put in nicer terms of course.

But the kicker here is that Jeff doesn't snipe or second guess. That's the important part. Too often leaders don't try to make it work, instead may actively try to sabotage their subordinate efforts just to prove that they were "right".

[Jaynit @jaynitx]

Jeff Bezos reveals the simple phrase that saved him countless arguments running Amazon

"Disagree and commit is a really important principle that saves a lot of arguing"

"One of my direct reports would want to do something. I'd think it was a bad idea. We'd go back and forth and I'd often say, you know what, I don't think you're right, but I'm going to gamble with you"

"You're closer to the ground truth than I am. I've known you for 20 years, you have great judgment"

"At least then you've made a decision and I'm agreeing to commit to that decision. I'm not going to be second guessing it, sniping at it, or saying I told you so"

"I'm going to try actively to help make sure it works. That's a really important teammate behavior"