Abraham Abel Boodala

Director of Innovation at Odyssey Therapeia, Bridging the brain with everything else. Neuroscience, Artificial Intelligence, and Biotechnology. Always learning.

[Abraham Abel Boodala @abrahamabel196]

@handsdiff Sounds very cool, excited to see and hear more from Precursor in the days to come

[Abraham Abel Boodala @abrahamabel196]

USA USA USA !!!! , my god what a team , what energy , the atmosphere

Best game I’ve seen of the World Cup so far!

[Abraham Abel Boodala @abrahamabel196]

@foodiedev_py @devgancode Check your dm 😂😂

[Abraham Abel Boodala @abrahamabel196]

@Ayussheth I’m in

[Abraham Abel Boodala @abrahamabel196]

@srijanshetty This is the way

[Abraham Abel Boodala @abrahamabel196]

@nabeelqu When my grandkids ask what life was like back then in 2026, I will say nothing, but wordlessly point to the lone frame atop my desk ,
the lone frame bearing those words so simple yet of depth unfathomable

"my Grok stupid
my Claude restricted
my GPT 5
Knicks in five"

[Abraham Abel Boodala @abrahamabel196]

Really sorry that you are going through this, I cant imagine how difficult it must be, And like so many kind strangers have said, it will get better, don't give up!, its probably easy for me to say that right now having no idea what you must be going through,

And in a world where theres so much you out of your control, the only thing you can do Is be kind to yourself, to be yourself, to explore create and do what makes you happy, its okay if things are not working out right now.

sharing a few things that helped me,

One was not benchmarking myself to anybody else is something that helps me , its difficult, but I've come to be at peace with what I am doing, and to go at life at my own pace,

Being kind to myself, more compassionate to myself, made me kinder to others, forgiving myself when fail at what I should succeeded at, being more understanding to myself when I make mistakes, being more kinder to myself when loose control of my emotions, doing things I wish could've done differently, is the thing that works the most for me right now and sometimes, just realizing that tomorrow is a new day just makes things easy

the best part of hitting rock bottom is that the only pace I can go , is up

Its difficult even now, but everyday is a challenge, and we deserve to be proud of ourselves for making it through one more day

Appreciating myself, and everyone around me, taking a genuine joy in other people, has kind of made me happier :) ,

Your courage in sharing your low is doing more than you can ever know for others, you've brought so many myriad strangers from every walk of life to stop by and express their support and share their experiences, and show their care, it makes me believe so much more in all of our shared humanity, it makes me hopeful for tomorrow

I genuinely wish and hope for your brighter tomorrow. It will get better , much better.

[Abraham Abel Boodala @abrahamabel196]

@sri9s probably because you'll get downgraded to opus 4.8 5x more often before you get to use Fable xD.
Sorry I'm just annoyed at the horrible guardrails, but seriously I've been wondering the same , good question!

[Abraham Abel Boodala @abrahamabel196]

quite unhappy right now, phenomenal model, horrible horrible guardrails (it literally downgrades every conversation to opus 4.8, at the start of) , will not pay for it on api, its too expensive, if available on the max 200$ , subscription then yes. Lets see how things pan out in 2 weeks.

[Abraham Abel Boodala @abrahamabel196]

@minchoi Its amazing yes, if one can get it to not downgrade every prompt to opus 4.8, so frustrating, like no matter what the context is, it keeps downgrading from fable 5. smh

[Abraham Abel Boodala @abrahamabel196]

Frustrating experience with Claude Fable 5 right now, no matter what I prompt, even the most harmless of questions, like help drafting an email reply it downgrades to opus 4.8, which is a sharp contrast to the experience I had earlier today where it one shotted the most complex of coding tasks. I Hope Anthropic are looking into this.

[Abraham Abel Boodala @abrahamabel196]

Siri AI is quite nice , excited to see if we could potentially sync with agentic workflows from other ai

#WWDC #WWDC26

[Abraham Abel Boodala @abrahamabel196]

came here to comment about this, its pretty but it really kills the experience, Im sorry to ask but , its so hard for me to go through the labs page right now, its slow , and not user friendly, I literally have no way of telling which experiments released when, cant see everything at once, the only way to search is via carousel, and its a little frustrating, because I have to repeatedly click the next arrow button , please bring the old ui back for this, or requesting more user friendly experience here

[Abraham Abel Boodala @abrahamabel196]

its here, all hail @OpenAI

[Abraham Abel Boodala @abrahamabel196]

@OpenAIDevs GPT image 2 to the rescue !

[Abraham Abel Boodala @abrahamabel196]

GPT 5.5 , amaze amaze amaze

[OpenAI Developers @OpenAIDevs]

NS41

[Abraham Abel Boodala @abrahamabel196]

@AnthropicAI @claudeaiOpus
Claude Opus 4.7 adaptive thinking feedback,
from a Max user on the $200 tier:I'm running into serious quality issues on creative and strategic chat tasks. No matter how many times I prompt "think deeply" or "ultrathink," the model refuses to engage its thinking most of the time, and even when it does, the responses are substantially worse than Opus 4.6 or 4.5 with extended thinking.

It feels like the adaptive router is consistently misclassifying non-code work as "low effort" and allocating almost no reasoning to it.For coding it works well, because I can set the thinking quality through effort levels. But in Claude chat, the removal of a manual thinking selector is a meaningful downgrade. As a Max subscriber, not being given the option to force deeper reasoning is genuinely hurting the product experience for me.

Concrete example from tonight: I spent 90 minutes on a multi-constraint strategy problem with 4.7, repeatedly adding context and re-prompting to try to get a thoughtful response. Switched to Opus 4.6 with extended thinking and it one-shot a genuinely useful answer.

For me the ask is simple: bring back a manual thinking selector in Claude chat, or expose effort levels to chat users the way the API does. The adaptive router isn't yet calibrated for non-technical complexity, and power users need the override while you tune it.

[Abraham Abel Boodala @abrahamabel196]

@hiitsmemooneh I’m so sorry for your loss

[Abraham Abel Boodala @abrahamabel196]

A truly sad night , more than 88 dead , and scores more injured , please stay safe Israel, I hope this won’t be long.

[Eli Afriat 🇮🇱 @EliAfriatISR]

Sad night in Israel.
Pray.💔

[Abraham Abel Boodala @abrahamabel196]

@SAlwashahi Praying for the safety of Dubai and the UAE , nothing will ever stop its Innovation and drive. 🇦🇪