Ferhat Yildiz

Owner at Turing Technology Services, Independent IT Contractor

[Ferhat Yildiz @turingts]

@r0ck3t23 I think this is a great benchmark for AGI. It cuts through all kinds of hype and noise.

[Ferhat Yildiz @turingts]

@jonobelotti_IO In retrospect, canary releases for the Enactor plan could’ve helped and fault injection could’ve discovered the race condition.

[Ferhat Yildiz @turingts]

@joaompinto4 @emollick @ollama Yea I don’t think it’s practical right now with at least 800GB RAM requirement for DeepSeek R1 671b. Some folks ran it with a cluster of 7xM4 Pro Mac minis but I don’t see a lot of people doing that. Either models need to shrink in size or the hardware needs to catch-up.

[Ferhat Yildiz @turingts]

@emollick @simonw Inference costs have increased though due to “reasoning” models such as o1 and o1-pro.

[Ferhat Yildiz @turingts]

@emollick @MTabarrok I think AGIs have to be competitively priced for them to replace any kind of labor. The compute costs and cost per task need to go down significantly. I don’t see this happening in the near term if I look at simple tasks on ARC AGI benchmarks causing enormous compute costs.

[Ferhat Yildiz @turingts]

@polynoamial 💯

[Ferhat Yildiz @turingts]

@Sauers_ ARC-AGI (v1) contains challenges which are easy for humans to solve according to @fchollet

v2 will be more difficult for o3 but still easy for humans.

I think we need a fundamental hardware change if we want to get to AGI.

[Ferhat Yildiz @turingts]

@MIT_CSAIL I think it’s amazing. $200/month for o1-pro is cheap judging by the kinds of problems it’s able to solve for engineers which would otherwise take many hours to solve.

[Ferhat Yildiz @turingts]

@eshear 💯

IMO it should be easy to understand and there should be less gatekeeping for pre-prints. I think the Open Science movement is pushing for this right now.

[Ferhat Yildiz @turingts]

@francoisfleuret @sama wrote the following thing. “It is possible that we will have superintelligence in a few thousand days (!); it may take longer, but I’m confident we’ll get there.”

[Ferhat Yildiz @turingts]

@pesterhazy @davefarley77 I think shipping unfinished features to production in a reliable manner requires feature toggles and a mature test suite.

[Ferhat Yildiz @turingts]

@pesterhazy @davefarley77 It’s very difficult to embrace unfinished features on main if production needs an urgent fix and deploying unfinished features to production leads to more trouble.

[Ferhat Yildiz @turingts]

@TomerUllman Exciting. Pre-Ordered 🙏🏼

[Ferhat Yildiz @turingts]

@kelseyhightower Recently noticed 25% higher performance and 30% cheaper compute on AWS Graviton3 (ARM) vs Intel based chips. This can make a big difference at scale.

[Guillermo Rauch @rauchg]

My take: software and networks are extremely, mindbogglingly complicated.

Turns out Kubernetes is a tool designed to model all of that complexity and make it deployable. It's astonishingly good at it.

When people beat on Kubernetes, they're either trolling or failing to appreciate the staggering amount of complexity it was designed and evolved to model.

The only intellectually honest path out of that complexity is to simplify our software stack. This is why you can't ship "good DX for Kubernetes", incidentally, and why many have tried and failed.

We learned this lesson early days of @vercel and why we refined our focus onto web frameworks and interactive frontend compute. It's no longer possible to deploy a Docker containerized 128gb mem Java backend monolith with 3 instances connected to a self-hosted database replica set with custom networks and upgrade policies.

It turned out making certain things impossible and saying "no" was the key to making a lot of developers happy and removing a lot of yaml config and devops headache from the world.

[Ferhat Yildiz @turingts]

@rauchg Having navigated the pre-Kubernetes era, I really appreciate how it has simplified a lot of things. It's been a notable shift in managing infrastructure and abstracting it away via APIs. Managed Kubernetes offerings by Cloud providers make it very easy to run Kubernetes.

[Marc Andreessen 🇺🇸 @pmarca]

THE TECHNO-OPTIMIST MANIFESTO part 1

“You live in a deranged age — more deranged than usual, because despite great scientific and technological advances, man has not the faintest idea of who he is or what he is doing.”
— Walker Percy

“Our species is 300,000 years old. For the first 290,000 years, we were foragers, subsisting in a way that’s still observable among the Bushmen of the Kalahari and the Sentinelese of the Andaman Islands. Even after Homo Sapiens embraced agriculture, progress was painfully slow. A person born in Sumer in 4,000BC would find the resources, work, and technology available in England at the time of the Norman Conquest or in the Aztec Empire at the time of Columbus quite familiar. Then, beginning in the 18th Century, many people’s standard of living skyrocketed. What brought about this dramatic improvement, and why?”
— Marian Tupy

“There’s a way to do it better. Find it.”
— Thomas Edison

Lies

We are being lied to.

We are told that technology takes our jobs, reduces our wages, increases inequality, threatens our health, ruins the environment, degrades our society, corrupts our children, impairs our humanity, threatens our future, and is ever on the verge of ruining everything.

We are told to be angry, bitter, and resentful about technology.

We are told to be pessimistic.

The myth of Prometheus – in various updated forms like Frankenstein, Oppenheimer, and Terminator – haunts our nightmares.

We are told to denounce our birthright – our intelligence, our control over nature, our ability to build a better world.

We are told to be miserable about the future.

[Ferhat Yildiz @turingts]

@veteran_peter @gdb I was able to pay for it using a credit card. Not sure whether debit cards work.