Sameh

@quntaai founder, SW engineer, AI engineer

[Sameh @samehalyassin]

@ThePrimeagen and it's demonstrated the real AI we all need and waiting for.

[Sameh @samehalyassin]

@mfranz_on the last episode and the missing flash memory 😂

[Sameh @samehalyassin]

@GergelyOrosz trying to take the biggest piece of the cake

https://t.co/iX6nvexW6R

[Sameh @samehalyassin]

there is a direction all AI big providores like Antropic and OpenAI are following, which is overwalmed the market with all use cases and potintial product applied over thier main platform, take in mind this not new in SW industry, it was almost (from Agile perspective) each sprint, but the differnace here is the hype for each feature they release even there is no real matrix to make sure the client will use it and keep using it, at the end they have may advantage by following this direction like they keep making hype and at the same time the hype makes them release more, and its a benchmark for the investor to tell them look we eating the market beside mony other benfits

[Sameh @samehalyassin]

@Yuchenj_UW still there is a hope to not give up for anthropic and openAi

[Sameh @samehalyassin]

@ahmd3ssam ومصطفى سليمان بميكروسوف في سبات بعد ما اصلا تم تهميشه ب alphafold

[Sameh @samehalyassin]

@julien_c @etnshow good one 🤣

[Sameh @samehalyassin]

@BashmohandesM واظن openAI عندهم هذا التوجه وفعليا بدأو بامور ال machines لكن على مستوى اقل من الروبوتات .

[Sameh @samehalyassin]

@ThePrimeagen the opposite of the Software engineering we used to know

[Sameh @samehalyassin]

@ThePrimeagen so we will hear them yelling again: "we told you "

[Sameh @samehalyassin]

@chddaniel it will be the best marketing that happens for all Anthropic models

[Sameh @samehalyassin]

@tickswiz @GoogleResearch not explainable, not predictable, not traceable.
no one's should run llm generated code directly in his data even inside cloned sandbox.

[Sameh @samehalyassin]

@GoogleResearch not explainable, not predictable, not traceable.
no one's should run llm generated code directly in his data even inside cloned sandbox.
in quntaAI we found new better approach can solve all above concerns and it's the less token burning.

[Sameh @samehalyassin]

@mfranz_on whats about the speaker 💔

[Sameh @samehalyassin]

@MatthewBerman

[Sameh @samehalyassin]

@theo is there something personal between you and the model? it try to make a fight with you

[Sameh @samehalyassin]

@yousefrol هذا الاسبوع كان نار بشكل عام مودلات اوبن سورس من انفيديا وغيرهم واليوم المجال كلو نشط .اظن نحن باحسن ايام ال ai

[Sameh @samehalyassin]

@mfranz_on do you recommend keep learning low levels like openCV for computer vision?
i am really passionate about CV in general but the recent AI hype made me lost

[Sameh @samehalyassin]

I have too much to say but it's even harder and more confusing for us to figure out how to work and learn,
I am trying to learn some new tools and code libraries, but honestly, I fail every time. Even the way we learned programming and SW is not valid and not fun anymore.
it's was easy before, just watch courses then dive in details and documentation . but now i am not sure if this still worth it.

[Sameh @samehalyassin]

The majority of what we see of AI applications, including agents, is not more than developers playing around and burning tokens pushed by FOMO.
The rest of the real solutions and products are still internal and for specific use cases