After adding an adversarial review gate to implementation plans and code I saw large uptick in quality. I use Opus 4.8 as plan writer and orchestrator. For adversarial reviewer I use GPT 5.5.
I still find things to tweak and fix up but the amount dropped pretty dramatically. As always I am responsible for what I ship so I review and test everything of course. I still think we are a ways away from fully automated software forge but what is currently possible is pretty cool.
S-1 isn’t public yet. Source on the lockup period? SpaceX for example filed with accelerated release of insider/investor shares so I don’t think we can know if this is the case until the filing documents become public.
Look at SpaceXs filing. There is one but it is super short. I was just pointing out that 365day lockup is likely incorrect and OP doesn’t really know that until the filing is approved and becomes public.
Going to give the benefit of the doubt here. I know what lockup period means.
365day lockup isn’t a universal standard. For example for SpaceX 20% of insider shares can be sold in the first few days. 100% within the first 3 months.
Without a public S-1 filing we don’t know what the lockup for Anthropic will be
Total market indexes and target date funds will include this and SpaceX on float adjusted basis I believe. The blast radius is much larger than funds that track the NASDAQ directly.
But isn't that what "total market" means? I don't see how if you invest in a total market fund you could declare "except for SpaceX, Anthropic and OpenAI". Why is it so bad for these accounts to be invested in these companies anyway? Seems pretty typical, i bet all kinds of companies are added to total market indexes each year.
Until recently companies that IPOed weren’t immediately added to the major indexes so there was a longer period for price discovery. This year that changed; so you have retirement funds that typically are more conservative acting as exit liquidity for these massive IPOs.
I would have less of an issue if the inclusion in major indexes was delayed 6-12months but we are looking at inclusion within like 5 days for some of these indexes.
The float will get bigger as you wait tho, since it's common for early investors to be locked for e.g. 6 months. You can argue it's better to smooth the entry as float gets unlocked rather than being front run by all the hedge funds in a single day on a massive capitalization.
Still the wrong response. Discipline the people abusing your trust.
Part of being a good manager is knowing when to step in and have a private conversation with people before things get too bad. If the bad behavior continues then you follow the process of formal write up’s and eventually termination.
Collective punishment is the sign of a manager who doesn’t know what they are doing and will kill a team.
There are plenty of EU domiciled Managed Service Providers who do have the skills though.
Having your government infrastructure run in country and managed by your citizens seems like a good idea just in general. It helps to develop local skills and the people living in country have a better feel for the needs of the local people.
I am an American but this just seems like a good idea even if the current geopolitical situation was better.
There are European cloud providers though. American hyper scalers are not the only option. Lidl, Hetzner, OVH, ..etc.
I can spin up a dedicated server within 24-48hours or a VM within minutes on OVH. Also there have been plenty of white papers written about how much more expensive AWS is when compared to Hetzner or OVH.
The big cloud providers are quite expensive and come with a lot of geopolitical risk/baggage. European governments have safer alternatives within their own borders.
Edit: it’s Lidl who launched a cloud service not Aldi
Specs weren’t the problem with waterfall. The difficulty in changing them to match reality was.
The waterfall process I experienced went like this:
- Product folks created requirements
- architects produced detailed specs
- project managers created tickets based on specs
- lengthy estimation ensued.
- Then finally developers proceeded with implementation.
- QA tested it.
Each step above involved lengthy review with like 5-10people. If the devs found an issue with the spec or god forbid the requirement it triggered a massive cascade of work for everyone above. Things needed to be reviewed again, customers may need to get contacted, …etc.
I think we can learn from that and optimize for change. Specs as living documents close to the code should be less cumbersome. But, just like anything else large corporations will probably fumble this like they did with “agile” (SAFe I am looking at you).
This is a long way to say specs aren’t bad. Specs that are difficult to change are though.
Changing the default behavior for all of your users with no notification is pretty unforgivable. Even if this feature worked correctly, it obviously doesn’t, this should at minimum be a prompt after upgrade to let the user confirm that this is what they want. But honestly should be opt in for those that want it.
To have it silently just start adding marketing copy to git commit messages is pretty bad. To have that added text not be visible to the user in the UI so they can remove it before commit is just much worse.
This kind of thing being released speaks to a greater disfunction over there. Not a good look at all and I am not a Microsoft or AI hater. But my commit messages are not where you move fast and break things
> Changing the default behavior for all of your users with no notification is pretty unforgivable.
I noticed that as soon as you make a bug report/feature request on VSCode's repo, you instantly get someone's OpenClaw agent with an automated pull request that sometimes wants to change defaults in the main codebase
Looks like AI is really trigger-happy with that, with zero understanding or care that there's thousands of users affected and it's not just one individual's settings.json
Also, the hallucinated PR does not necessarily address the original issue whatsoever, just like this PR. It should have functionality to detect AI-authored code, but whoever made the PR skipped actually doing the hard work and just changed a default to always on, exactly the kind of misunderstanding you see with OpenClaw shotgun PRs
And then they apparently posted an alibi "I'm sorry" here. Or maybe it is genuine, but the choice is between incompetence and fake "I'm sorry". Where is QA?
> To have it silently just start adding marketing copy to git commit messages is pretty bad. To have that added text not be visible to the user in the UI so they can remove it before commit is just much worse.
This is one of the problems, but it is not only one. To be better, should be:
1. It should be visible in the UI for entering the commit message, to make it clear what it is doing.
2. It should not add such a thing if the Copilot is disabled. (It is mentioned by dmitriv and would hopefully be fixed soon enough)
I do not use Copilot nor any other LLMs nor VS Code, but if the problems are corrected then I think the feature would probably be reasonable.
No, it's fine. I really hope that more people will switch to something else, like Neovim, Emacs, or any other open-source editor where such unacceptable situations are practically impossible. I hope more people will start to value their privacy and right to choose, and find the courage to say gtfo and switch to something else. Because this is unforgivable.
It just means that when changing a global default with such impact the user should be prompted with an option to opt out of the new behavior. Something like “AI assisted changes will now have ‘coauthored by Copilot’ added to the commit message”. If the user clicks “no thanks” it changes their local setting to “off” to opt them out of this new global default.
Not sure of what current prices look like but an old desktop sitting on the floor of your office might work well for you. You would need decent internet but running a single node kubernetes cluster as a GitHub action runner has worked well for others I know.
A buddy of mine runs his whole CICD setup off an old gaming desktop. They use tailscale to connect to their hosted infrastructure and set it up as a GitHub action runner.
reply