GovTech

Originally Posted by Unregistered

Dsaid also a bit awkward one. Develop the launchpad then bank hard on it.

The recent dsaid conference also damn small. Should bring more ppl

That being said, the venue and food was good

I'm still laughing at the launchpad vs pair comparison. Unfortunately while launchpad had a good roadmap, their execution and end result lacks the polish and hence shitty. Too many use cases wont work with their product. And what happened to pair now? Announce big big on OGP copilot in newspaper, then crickets until now. End up we just wait for o365 copilot to come.

Originally Posted by Unregistered

I'm still laughing at the launchpad vs pair comparison. Unfortunately while launchpad had a good roadmap, their execution and end result lacks the polish and hence shitty. Too many use cases wont work with their product. And what happened to pair now? Announce big big on OGP copilot in newspaper, then crickets until now. End up we just wait for o365 copilot to come.

OGP got Pair, DSAID got LaunchPad, and what happened to GDS LLMStack? Every division wants to develop similar LLM products, but in the end, all no news already.

case of too many cooks spoil the dish?

The dish being yet another wrapper around openai API?

No wonder govtech needs to reduce head count.

Originally Posted by Unregistered

The dish being yet another wrapper around openai API?

No wonder govtech needs to reduce head count.

But has agreement with Microsoft not to use our data for training purpose

cut the deadwood away

Originally Posted by Unregistered

cut the deadwood away

Say properly hor, who is the deadwood? Corporate? GDS? Dsaid? Ogp? Csg? Who else

Originally Posted by Unregistered

Say properly hor, who is the deadwood? Corporate? GDS? Dsaid? Ogp? Csg? Who else

What do you think?

Originally Posted by Unregistered

What do you think?

I'd say chip every here and there...

Maybe that's why contract is the first to go

Originally Posted by Unregistered

Been there done that multiple times(with success). Unfortunately im just a lowly person and the most i can influence the projects around me, and this would require higher ups to evolve(lol in public service). For now I will just take it one project at a time. Maybe one day
I will just follow the rest and lie flat, or maybe one day i will just quit quietly. Who knows?

And that's the exact problem. I've also done heroics during my time, but that was not repeatable over the long run (not to my mental health, and also not sustainable in terms of required talent).

If i've stayed on, I'd get embroiled in arguments with people on top who know nothing about the subject but insist that they're right. (as evidenced by a similarly talented colleague who decided to stay on)

Originally Posted by Unregistered

No one asking for insource every single project. Outsource those that are low on rewards, and insource those that are highly valuable. Just getting a small team of 2-4 SWE for every agency could do wonders in supporting the agencies.

It should be outsource when the system is stable and requirements not in a flux. Insource when nobody know what they wants and we are in a pathfinding stage. It's not about high or low value.

Originally Posted by Unregistered

If outsourced SI can do it cheaper and better than govtech, then govtech has failed.

No they can do it cheaper than govt if the system is not changing and crashing all the time.

I would even go on to say the system after the pathfinding stage should be reimplemented by commercial (after the requirements are firmed up) for maintenance in mind.

Originally Posted by Unregistered

We are not R&D, and we dont produce products competing with commercial.

Pathfinding is R&D, and that should what govtech should be doing. no commercial entity would touch govt r+d unless they are paid obscene amounts because the contract would be so open ended.