@DmitryDoreshkin@Jason She is not very friendly to tech.,,meaning I had to give several big talks to her on why she needs to use it. Now I’m trying to get her to adobt AI bots and learn how to use the one I host at home for her and the rest of my family,
Your 1% is low…gotta get those numbers up.
@Jason I got a Cybertruck because it could drive for me and Its one of the safest vehicles on the planet. I also forced my wife to use fsd on her model Y for the safety of our children. If you wanna shit on FSD you should have to show your stats.
@bcherny Most of your posts I see are just you responding to complaints to post chat numbers so you can reference it…its almost like yall blindly make changes without knowing whats actually happening.
In my uneducated opinion yall lost the plot when you cut off openclaw and similar applications…meanwhile every other AI provider leaned in. Yall may have one of the leading models for now, but going from darling to villain in 90 days is not a good look. I’m not sure who in the boardroom said this was a good idea, but from my perspective yall are getting close to being very behind because a lack of leadership and foresight. Dont lean on Mythos, everyone else will have that soon too….hell china will release the open weights in a month,
The way you treat your users today is going to have a lasting effect.
@steipete@steipete have you posted your workspace anywhere? At least once a day Im wondering how Molty is setup and configured vs my bot. My bot has all of this, but would never be so proactive in getting something like this done.
The problem it solves and the initial price tag, what’s wrong with that today? How far will that go down when it’s a more popular service. Your just seeing people trying to get to the market first and set price standards. Commercial service companies do 100k+ deals every week. What’s 20k to them to get a company that is trying to stay ahead in AI, and can help implement workflows to prevent leakage? Large companies is even crazier, I’ve seen 100k+. Early bird gets the room. These things take a lot more time then X will lead you to believe, its going to get quicker and easier as time goes on. The companies doing it for 20k today, will be the 1hr setups for 5-10k next year and will be able to stay competitive with the market flood of AI implementation companies.
@Voxyz_ai So what’s different in telegram/openclaw then what it has been? I’ve had STT and TTS via elevenlabs for months now. Some of my older users prefer just using voice. Sending a voice message would trigger a voice message received from your bot in return.
Key thing your missing for GB is Dallas was the worst Defense for a whole season. Looks a lot better now, but we also lost an entire season of everyone's prime because of this. I like the return, but sometimes things like this get lost. If was not fun for a Cowboys Fan this past December.
I have tried codex on my mac, codex cli, my openclaw bot. I haven’t tried chatgpt, didn’t think that was a good route.
Prior to the opus switch i could just have a bot build a pretty nice website to my standards that wasn’t slop. And thats just basically going back and forth with cloudflare preview links and giving my notes and thoughts. 2 hours later, I have something I like.
I figured it may just be the way llm is injesting the prompts with opus vs gpt5.4.
How can I make gpt5.4 give me better output? Any config or prompting suggestions? Its my main LLM now.
I had to go to claude design today to build out a website, gpt5.4 would have never done this. It has just been giving me ai slop. I think its just claude gets us vibe coders better output, or we I just cant prompt gpt5.4 correctly? I would love to know how yall are getting the success yall are.