Speaking in an interview with the Financial Times, Suleyman talked about "professional-grade AGI" and how Microsoft expects it to capture a large share of the enterprise market.
Until the AI results can be trusted, I don’t see how this happens. I’ve been using AI for some questions that would normally be on stackoverflow but I don’t find code generation to save me time. Because I can’t implicitly trust the product, I still have to review the code before I can use it. If I have to review and understand it, it rarely saves me time. There have been edge cases where it helped me in some areas, like turning a CSV into a visual report in PDF format but I still had to review everything. It just happens that I suck as report tools so it was a shorter amount of time for me to review the AI report than to put together visualizations myself.
I’d offer a small correction: that ain’t happening as long as companies are liable for the AI’s work. If companies can just blame the model and get away with a fine that’s less than the savings, they absolutely will take that deal.
Keep companies accountable and the bubble will burst
You’re not using it correctly. You’re supposed to vibecode the entire application by defining good parameters.
You don’t debug or fix stuff, you just iterate. You just make a new application with revised parameters.
If you tell the LLM “this is bad, make it better”, it will have the bad thing im it’s context and it will therefor try to make the bad thing again.
Instead, if it makes a mistake,you throw out the whole thing and start over witg revised parameters.
This will save us money in the short run. In the long run… who cares.
if you tell the LLM “this is bad, make it better”, it will have the bad thing im it’s context and it will therefor try to make the bad thing again.
You forgot “/s” I tried that a few times. With and without welling what’s wrong. After 3-5 times it gives you the first solution it offered. Tell them that and it ignores it.
You can’t trust that it’s impossible by it’s architecture like if you tell it reset your memory… and it will simulate that it forgot, but it didn’t and it will affect all prompts
This is way all models easily leak their system prompts.
Until the AI results can be trusted, I don’t see how this happens. I’ve been using AI for some questions that would normally be on stackoverflow but I don’t find code generation to save me time. Because I can’t implicitly trust the product, I still have to review the code before I can use it. If I have to review and understand it, it rarely saves me time. There have been edge cases where it helped me in some areas, like turning a CSV into a visual report in PDF format but I still had to review everything. It just happens that I suck as report tools so it was a shorter amount of time for me to review the AI report than to put together visualizations myself.
I’d offer a small correction: that ain’t happening as long as companies are liable for the AI’s work. If companies can just blame the model and get away with a fine that’s less than the savings, they absolutely will take that deal. Keep companies accountable and the bubble will burst
You’re not using it correctly. You’re supposed to vibecode the entire application by defining good parameters. You don’t debug or fix stuff, you just iterate. You just make a new application with revised parameters.
If you tell the LLM “this is bad, make it better”, it will have the bad thing im it’s context and it will therefor try to make the bad thing again.
Instead, if it makes a mistake,you throw out the whole thing and start over witg revised parameters.
This will save us money in the short run. In the long run… who cares.
You forgot “/s” I tried that a few times. With and without welling what’s wrong. After 3-5 times it gives you the first solution it offered. Tell them that and it ignores it.
You can’t trust that it’s impossible by it’s architecture like if you tell it reset your memory… and it will simulate that it forgot, but it didn’t and it will affect all prompts
This is way all models easily leak their system prompts.