Tactics/os
@Tacticsos
Followers
477
Following
5K
Media
8
Statuses
3K
Joined March 2017
RT @MacaesBruno: Jordan is now very surprised to find out Israel considers Jordan to be Israeli territory. I don’t know any other governmen….
0
51
0
RT @jeremyphoward: Spent the last couple of days trying to do a lot with GPT-5 on the chatgpt web app. Sorry to say I'm giving up on it :(….
0
55
0
RT @Birdyword: Remember slightly over three weeks ago when this was so awe-inducing that he compared it to the Manhattan Project.
0
539
0
RT @peligrietzer: Imagine being Paul Graham and gradually finding out you are the only person in your moral subculture (SV libertarians) wh….
0
18
0
RT @MouinRabbani: It is unclear whether Fox is calling for the Israeli air force to kill all of al-Sharif’s remaining colleagues and bomb A….
0
33
0
RT @fchollet: We were able to reproduce the strong findings of the HRM paper on ARC-AGI-1. Further, we ran a series of ablation experiment….
0
290
0
RT @ClementDelangue: Don't trust all the social media hype of AI products, a lot of it is hidden paid (this one from a popular AI tool) htt….
0
30
0
RT @NoahCarl90: Life expectancy over time in Gaza, along with four other countries that experienced wars in recent decades. .
0
2K
0
RT @JehadAbusalim: People do not grasp the scale of the reactionary measures Israel and its supporters are unleashing to salvage their geno….
0
2K
0
RT @academic_la: When people say the problem is Netanyahu and not the Israeli people, I suggest they look at the polls. Likud is still the….
0
1K
0
RT @yuval_abraham: אחרי השבעה באוקטובר באמ"ן הוקם צוות שכונה "תא לגיטימציה". אנשי מודיעין שחיפשו מידע שיעזור להעניק "לגיטימציה" למעשי הצבא….
0
850
0
I can tell you too that OAI are not the only ones who built or played with a model router. xAI did it. The LMArena tested one. Switchpoint Router is a thing. I think Google tested one(?). OAI are just unique in making stripping control a "selling feature.".
The model router becomes a pet idea among executives and they decide to scrap the o3 release in favor of "GPT-5.".6. They've having more struggles with o3 in general then they're willing to admit and the router idea doesn't really work well (but it's a pet idea, so. ) (4/x).
2
0
15
If they had delivered something GPT-OSS style (benchmaxed model of limited broad utility), they'd be in a far worse circumstance. That might be company-killing.
5
1
179
For people voting "six months" or a "year," six months brings us back to 3.7 Sonnet being the strongest Western model. A year brings us back to 3.5 Sonnet (not 3.6, 3.5) being the strongest.
0
0
3
Just to be clear, I don't think GPT-5 is some sort of awful model Llama 4 style, I just explaining why things have gone so poorly despite it being at the very least solid.
1
4
217
RT @Isinlor: People don't talk enough that ClosedAI and the closed science that followed across labs slowed down progress in AI significant….
0
4
0
How far do you think the top Chinese AI labs are behind the top Western "frontier" labs?.
2
0
3
RT @jeremyphoward: Now that the era of the scaling "law" is coming to a close, I guess every lab will have their Llama 4 moment. Grok had….
0
106
0