Well since you work at a lab you should know that most capabilities arise in pretraining, not posttraining or mid training, and the latter two mostly function to bring out the hidden intelligence in these models more than anything else.
It hasn't worked for nuclear disarmament. We live in a world where many countries have nuclear arsenals. "But it hasn't killed us yet!" Yeah sure, it's only been less than a century since they were invented. Who knows when nuclear war will come?
Sensationalist garbage. Actual studies have found that the loss in revenue is minimal, and that current wealth taxes are well below what they should be for maximum benefit:
> We show that trickle-down effects do exist, but that they are quantitatively small. A one percentage point increase in the top wealth tax rate decreases aggregate employment by 0.02%, aggregate investment by 0.07%, and aggregate value-added by 0.10% in the long run. Importantly, these effects are modest despite the fact that top wealth holders—many of whom are entrepreneurs—account for a large share of economic activity in Scandinavia through the businesses they control. Our approach to estimating trickle-down effects is arguably the most innovative part of our paper. It is based on clear identification assumptions and is statistically precise.
> The modest economic effects of tax-induced migration do not necessarily imply that wealth taxation is an optimal policy. To evaluate wealth taxation, we also have to account for their effects along the intensive margin, operating through changes in savings, investments, avoidance, and evasion. Jakobsen, Jakobsen, Kleven and Zucman (2020) find sizable intensive margin effects of wealth tax reform in Denmark. Combining the migration estimates presented here with their intensive margin estimates, we show that the Scandinavian wealth taxes were below the Laffer point and that their Marginal Cost of Public Funds (MCPF) was about 4.2.54 Leaving aside equity arguments, taxing top wealth would be welfare-improving if the revenue raised is spent on projects with a Marginal Value of Public Funds (MVPF) greater than 4.2. Comparing MVPFs across a range of policies, Hendren and Sprung-Keyser (2020) argue that programs targeted to low-income children have the highest MVPFs, often greater than 5. This suggests that funding projects for low-income children via progressive wealth taxation has the potential to increase social welfare.
Source: also work at a lab.