Sam 🫩

Locked In.

[Sam 🫩 @samuelayibatari]

20's - billionaire
30's - trillionaire

[Sam 🫩 @samuelayibatari]

It seems @paystack devs are Attack on Titan fans 😂

[Sam 🫩 @samuelayibatari]

@ctatedev @rauchg Won't you still need humans to learn it so as to generate quality training data to train the models?

[Sam 🫩 @samuelayibatari]

@elite_developer A custom operating system built with a custom programming language that was built with c++.

[Sam 🫩 @samuelayibatari]

@sakagramming Haven't people heard of react native, TypeGPU, assembly script and so much more?

[Sam 🫩 @samuelayibatari]

@phl43 @ewinsberg @ShamikDasgupta1 Understanding is one such characteristic we are certain LLMs do not possess because, as rudimentary as it sounds, LLMs are fundamentally prediction machines, albeit really impressive ones. LLMs cannot mimic or emulate human intelligence.

[Sam 🫩 @samuelayibatari]

@phl43 @ewinsberg @ShamikDasgupta1 Intelligence is difficult to quantify because it lacks a definitive definition. So we evaluate intelligence based on its perceived characteristics not what it fundamentally is. Using these traits as isolated metrics is flawed since they are highly interdependent.

[Sam 🫩 @samuelayibatari]

@SpermCommander @Hugo71890282 @Sam278489329134 @sult @dalmaer Come on now, you didn't have to lie to prove your point...

[Sam 🫩 @samuelayibatari]

Check out Bizzo on Google playstore: https://t.co/WlkqcCOOV1

AppStore coming soon...

[Sam 🫩 @samuelayibatari]

I've been offline for some time, that was because school and this project I was working on. Introducing Bizzo, in short Bizzo is a p2p marketplace, designed for Nigerians by Nigerians, designed to provide real Nigerian sellers an audience with real buyers within their city.

[Sam 🫩 @samuelayibatari]

@QuanticASI @WaveTheoryUK That's not what he said, can't you read?

[Sam 🫩 @samuelayibatari]

@detransanon @d33v33d0 Because LLMs are fundamentally probabilistic prediction machines rather than reasoning engines, they can never efficiently navigate a strictly deterministic game, regardless of how much text-based game data they process.

[Sam 🫩 @samuelayibatari]

@detransanon @d33v33d0 Intelligence is a broad metaphysical concept that cannot be adequately represented by text alone. This limitation explains why LLMs cannot master chess.

[Sam 🫩 @samuelayibatari]

@burkov @broadfield_dev LLMs do not possess true comprehension or logical reasoning so these errors are to be expected, I believe they could be mitigated if the model was hooked up with a linter where it could cross reference its generations for any errors, this could also prevent hallucinations in code

[Sam 🫩 @samuelayibatari]

@burkov @broadfield_dev I have, quite a lot of times actually, aside from the occasional semantic inconsistencies, it's still prone to making some weird syntactic errors that it could have easily fixed if it had linters, mind you errors that even junior developers wouldn't make.

[Sam 🫩 @samuelayibatari]

@NebulaAI PS. <1% can afford group 3 right now.

[Sam 🫩 @samuelayibatari]

@RashedatJ I use my friend's mac 🙂

[Sam 🫩 @samuelayibatari]

@Joy678320811181 @CovingtonORINAS @fchollet It's so annoying at this point, it's becoming harder to use Twitter.