Dr. Bobby Gomez-Reino

4.6K posts

Dr. Bobby Gomez-Reino banner
Dr. Bobby Gomez-Reino

Dr. Bobby Gomez-Reino

@BobbyGRG

Physics PhD, entrepreneur, ex-CERN innovator, now CEO at Cleverdist. Fast-tracking value to industry with software & generative AI.

Switzerland انضم Nisan 2010
311 يتبع816 المتابعون
تغريدة مثبتة
Dr. Bobby Gomez-Reino
Dr. Bobby Gomez-Reino@BobbyGRG·
Small Teaser of our IO (Industrial Operator) 2.0 An AI system for Industrial Control Rooms. Version 2 going to production in early September. Version 3.0 (2025 Q4) in the making will use a new "Omniscient feature" we are developing based on hierarchical Realtime agents, where the industrial process intelligence is distributed across an intelligent tree. Testing now with the new @OpenAI gpt-realtime... Teaser coming soon too.
English
0
0
8
1.8K
Dr. Bobby Gomez-Reino
Dr. Bobby Gomez-Reino@BobbyGRG·
If model fusion feels like a revelation, then you've spent too much time treating LLMs as chatbots and not enough time treating them as reasoning components inside real systems.
English
0
0
0
5
未知
未知@luyun0120·
@BobbyGRG @jerryjliu0 说比做容易，指责比担责简单。这是大多数人的选择。 「For most real-world workflows, the cost-accuracy frontier ha...」
中文
1
0
0
7
Jerry Liu
Jerry Liu@jerryjliu0·
As much as I want to regain access to Fable, let's remember that 99.99%+ of the world has barely scratched the surface on GPT-5.5 and Opus 4.8, not to mention open-weight models 1. Very few people in F500 are tokenmaxxing effectively 2. For most real-world workflows, the cost-accuracy frontier has not been met
English
35
7
227
23K
⿻ Andrew Trask
⿻ Andrew Trask@iamtrask·
@AndrewCritchPhD it really is. imo - it's also why we're going to end up wtih "a network of neural networks" in the end. the mainframe AI era is ending.
English
1
0
5
586
Andrew Critch (🤖🩺🚀)
Andrew Critch (🤖🩺🚀)@AndrewCritchPhD·
Do people consider this research? This is a years-old perspective update for users of theMultiplicity and other Encultured multi-frontier workflows. Is it *still* surprising people that multi-agent workflows can outperform single agents on safety and robustness?
Andrew Critch (🤖🩺🚀) tweet media
English
4
1
14
1.8K
Dr. Bobby Gomez-Reino
Dr. Bobby Gomez-Reino@BobbyGRG·
@casper_hansen_ based model << generic harness/scaffolding << specific harness scaffolding I am confused with the fuzz. I think the main insights revealed from all this is the big gap between what llms can enable a what the world is really with them
English
0
0
0
169
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)@teortaxesTex·
I have tried to use OpenRouter Fusion API with cheap open models only, and saw reasoning that surpasses any of them individually. Then I looked into API logs and saw that this "Fusion" still calls Opus 4.8 as a judge. I see no way to disable it. Not cool, OpenRouter. Not cool.
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞) tweet mediaTeortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞) tweet media
OpenRouter@OpenRouter

Introducing the Fusion API, the smartest compound model in the market. Fusion achieves Fable-level intelligence at half the price. How it works 👇

English
31
16
509
47K
Dr. Bobby Gomez-Reino
Dr. Bobby Gomez-Reino@BobbyGRG·
@MichelIvan92347 true. i still think most people wont need (or know how to) use models of mythos+ skills. they will just feel the impact through services and applications they will use. I need fable level intelligence in our IO ( a - fashionably called now- "model fusion" powered system lol)
English
1
1
1
27
Dr. Bobby Gomez-Reino
Dr. Bobby Gomez-Reino@BobbyGRG·
i just saw this, and i realized that it's been less than 1 year since this IO 2.0 version 2.0 and it now looks soooo primitive compared to the current 4.4 one! need to share a video here soon
Dr. Bobby Gomez-Reino@BobbyGRG

Small Teaser of our IO (Industrial Operator) 2.0 An AI system for Industrial Control Rooms. Version 2 going to production in early September. Version 3.0 (2025 Q4) in the making will use a new "Omniscient feature" we are developing based on hierarchical Realtime agents, where the industrial process intelligence is distributed across an intelligent tree. Testing now with the new @OpenAI gpt-realtime... Teaser coming soon too.

English
0
0
3
96
Dr. Bobby Gomez-Reino
Dr. Bobby Gomez-Reino@BobbyGRG·
@rihim_s @viemccoy and the cost vs value only work for clearly identified use cases. not holistic "i want Ai to run my control room operations ". then mission by mission growing. also the cost for intelligence band keeps decreasing unlocking new missions 🙃
English
0
0
0
21
Dr. Bobby Gomez-Reino
Dr. Bobby Gomez-Reino@BobbyGRG·
@rihim_s @viemccoy sure. this cannot work for end customers effectively but we are doing it for industrial use cases. 3 system in prod (power plants, nuclear servicing factory and CERN CMS ..last one soon in control room). need to make the cost vs value maths well
English
1
0
1
48
𝚟𝚒𝚎 ⟢
𝚟𝚒𝚎 ⟢@viemccoy·
in college me and the professor I was working with did this with gpt-3.5-turbo and got gpt-4 level performance we never published it but Varibot remains forever vindicated.
OpenRouter@OpenRouter

How does it work? When you send a prompt to Fusion, we fan it out to a panel of models in parallel, each with web search and bash tools enabled. A judge model reads every response and extracts the structure: consensus points, contradictions, partial coverage, unique insights, blind spots. Chatroom: openrouter.ai/fusion

English
4
2
176
15.4K
Dr. Bobby Gomez-Reino
Dr. Bobby Gomez-Reino@BobbyGRG·
@viemccoy btw. many issues come after to figure out tricks to deal with latency for large trees or manage both asynch and synch work (self managed by the system)
English
0
0
0
37
Dr. Bobby Gomez-Reino
Dr. Bobby Gomez-Reino@BobbyGRG·
@viemccoy its been since 2024 like this. topology of conversations (we use dag and tree often, scratchpad tools to synch responses for large topologies, etc) . its been always obvious to us that this was always >> than base models used. feels so weird to see so many surprised.
English
1
0
1
235
Dr. Bobby Gomez-Reino
Dr. Bobby Gomez-Reino@BobbyGRG·
why r open router findings surprising for so many?. isnt this like 2024 late knowledge? please someone explain to me what am i missing!
English
0
0
0
41
Dr. Bobby Gomez-Reino
Dr. Bobby Gomez-Reino@BobbyGRG·
its not just about their salary. create a belivable European lab and they will come. CERN was built with that spirit and it worked. we dont have decision makers with the balls we used to. it should not be spread around Europe but centralized and scaled up. and i still think cern infra and site could be repurposed fast. then billions should be poured on, not countries (for reference many countries contributr 80-120M to cern... not enough, need big investors and banks... and they want a believable story. small note... we still need too a lot of chips
English
0
0
1
71
antirez
antirez@antirez·
What Europe should do right now: 1. Call all the European researchers working on AI and return them back with same salary (or they can stay but switch career). 2. Fill EU places having GPUs with money, and put those people there. 3. AI partnerships with China + India.
English
354
356
3.1K
191.1K
Dr. Bobby Gomez-Reino
Dr. Bobby Gomez-Reino@BobbyGRG·
@Youssofal_ yes! this morning i started some Cursor tasks and needed to pause and recalibrate after using fable heavily for days, because i knew 5.5 wouldn't take the whole job without breaking it and some baby sitting. 5.5 is an incrdible model... but we can't unsee what we've seen...
English
0
0
0
183
Youssof Altoukhi
Youssof Altoukhi@Youssofal_·
"Use 5.5, it is almost as good as Fable!"
Youssof Altoukhi tweet media
English
27
5
260
21.9K
Adam
Adam@_overment·
@BobbyGRG something’s up with the water, I guess
Adam tweet media
English
1
0
2
16

اكتشف

@luyun0120 @jerryjliu0 @iamtrask @AndrewCritchPhD @casper_hansen_ @teortaxesTex @MichelIvan92347 @rihim_s