Daniel

[Daniel @hoffmanndaniel]

@bundeskanzler Ich hoffe wirklich für mich und meine Familie, dass sie uns Bürgern einen #HowToSurviveAfD Guide hinterlassen, wenn sie brutal von der AfD mit absoluter Mehrheit kaltgestellt werden bei den nächsten Wahlen 🙈 So ein realitätsferner Mist 🙁

[Daniel @hoffmanndaniel]

@OpenAI Will try! At the moment my Plus limits with GPT 5.5 are burning faster then Claude Code with Fable 5 - it’s insane :-/

[Daniel @hoffmanndaniel]

@thsottiaux Sorry @thsottiaux but what would really boost productivity is if I didn't have to switch from Codex to Claude Code after two days just to keep working for the rest of the week :(

[Daniel @hoffmanndaniel]

@OpenAI Sorry guys but what would really boost productivity is if I didn't have to switch from Codex to Claude Code after two days just to keep working for the rest of the week :(

[Daniel @hoffmanndaniel]

@TimJayas Fully agree! Switched back from Codex to Claude code for some weeks now … It’s crazy, Codex is done after a few days a week. :(

[Daniel @hoffmanndaniel]

@amix3k But I have the huge issue that I can get 15min of excellence before I hit the Pro limit. Or half a day #Opus just like -25% of the excellence… my decision is clear.

[Daniel @hoffmanndaniel]

How quickly the wind changes… Switched from #ClaudeCode to #Codex because I hit the limits after 15min Opus. Now I’m back as can get max 15min GPT 5.5 out of the Pro subscription but like half a day with Opus 🙈

[Daniel @hoffmanndaniel]

@xpeten Bürokratie? Selbstzerstörung? Da gibts schon einiges…

[Daniel @hoffmanndaniel]

Same here! I silently switched back from #Codex to #Claude in the last days because I hit the limits in like one day. Surprised how well Claude is working right now + solid limits 👍

[BlockedPath @BlockedPaths]

I used Claude way more here in this instance than Codex and you can see the difference. I have a lot of usage left on Claude compared to Codex. Fishy

[Daniel @hoffmanndaniel]

@ulfposh Es gibt ja wohl wichtigere für die Regierung! Den Kiffern endlich wieder den gar ausmachen oder deutsche Unternehmer entlasten. Man kann nicht alles haben! 🤡 #CDUfährtDeutschlandandieWand

[Daniel @hoffmanndaniel]

Fully agree! Big fan of #Gemini 3 Flash in #openclaw as fast and cheap daily runner. 3.5 added weird stuff, was talking more then GPT back in the days and burned my budget - no thanks!

[Jeremy Howard @jeremyphoward]

Gemini Flash 3.5 is such a disappointing model.

It's intelligence and speed is awesome. Absolutely amazing.

But it's been trained to max evals, not to be helpful to humans.

It goes off and does random crap "for me" rather than just doing what I asked.

[Daniel @hoffmanndaniel]

@thsottiaux “The thing with the broken limits/token usage” 😬😢

[Daniel @hoffmanndaniel]

@alex_avoigt It’s an awesome car but I have my doubts especially Germans are willing to pay 99€+ months for auto steer (for those not interested in FSD) Sad move especially for me as first EU Model 3 owner and shareholder :-/

[Daniel @hoffmanndaniel]

@MMittermeier Noch ein paar wenige Jährchen „Brandmauer“ 🤡 und wir werden es leider alle erfahren 😒

[Daniel @hoffmanndaniel]

@BenjaminBadejo @humphreyCSGO Oh I was wondering! It’s 17,90€ in Germany 😬 a bit hefty

[Daniel @hoffmanndaniel]

@Jaaneek @cherry_mx_reds Can you tell how does it compare to the OpenAI subscription? I’m using GPT5.4 medium in a PA case openclaw and need to switch as it became so expensive and buggy in OC that I use 20% on my plus sub to create 2 appointments and read out an email :-/

[Daniel @hoffmanndaniel]

@thsottiaux You mean the heavy token burn the last days/weeks on 5.5? 😬

[Daniel @hoffmanndaniel]

@iAmHenryMascot I’ll try to give more details https://t.co/bDGlgBNVbj thanks!

[Daniel @hoffmanndaniel]

@Angaisb_ Horrible for me. Fixing for hours after every update :( Start right away with Hermes! It will save you a lot of time and nerves.

[Daniel @hoffmanndaniel]

@ClawiAi I’ll give up on openclaw - I want spending my time USING the bot , not fixing for hours after EVERY damn update :( Migration tips? Will it work better with Grok now than GPT?