You see part of the problem in gutting the General Theory in order to marry Keynesian Effective Demand with neoclassical economics in one sentence here:

Samuelson said the second half of the 20th century had reduced his hope that a purely Keynesian approach would be the economic salvation for nations. When governments become too big and controls too much of a nation's wealth, they become wasteful and blind to the needs of its citizens.

This is the post-WWII military Keynesianism, which follows from a misunderstanding of the monetary system and a shell game with the problem of uncertainty in which it is downgraded to mere calculable risk.

However, over a certain threshold size that most industrial democracies passed sometime pre-WWI, how big the central government is, is largely independent of whether the central government engages in active fiscal policies - especially in a federal system when the national government engages in the most effective automatic stabilizers.

One of the most effective automatic stabilizers that a national government can have in place are a job guarantee program, in which everyone who applies for work is allocated to a job at an established minimum living wage. That can be operated on a cadre system, in which a smaller supervisory staff adds to its ranks when there is an upsurge in JG applications out of the people with supervisory skills in the applicant pool. Since the wage is an entitlement, it can maintain a very high job/$ ratio by relying on labor intensive techniques and operating on tight labor on-costs. It can be further kept down in size as a permanent bureaucracy by permitting state and local governments to submit projects for JG labor, with project rotation and queuing encouraging state and local governments to use JG labor for wish-lists rather than to perform ongoing labor.

Complementing that is a program of automatic grants-in-aid to states/provinces when their tax revenues slump in response to an economic downturn to compensate $/$ for the lost tax revenue.

One could engage in quite effective Keynesian counter-cyclical policy using this approach, combined with a relatively small central government during times of economic expansion. Whether you should, or should not, depends on policy positions regarding the long term share of government investment and collective provision of social goods desired. However, the more direct Keynesian intervention is perfectly compatible with a much smaller central government than currently prevails ... and smaller than the size of government required for stability under hydraulic Samuelsonian Keynesianism ... where the fear of the Phantom Menace that is the imagined impact of deficits on interest rates in the middle of recession leads to much heavier reliance on the Balanced Budget multiplier, which requires a large government share of GDP as a stabilizer.

Of course, post-WWII, y'all got not so much military in that larger government share and more collective provision of social goods. Perhaps for some reason the collective memory of WWII was not so fond, even among the winners. But in the politics of the US, it was the military where hydraulic Keynesians could find the avenue to maintain a larger government share of US GDP.

One of the reasons that Samuelson kept losing to Friedman in debate is he conceded too much ground to a mindless Monetarism by working within a framework of exogenous, government determined money supplies, rather than exogenous, government-determined, price of liquidity. Government has repeatedly demonstrated its capacity to fix the price of liquidity at any desired level from 0% on up, and repeatedly demonstrated its inability to regulate the quantity of money in the economy, but in teaching legions of students, via his textbooks and rival textbooks that have to have 90% identical material to be adopted, in their required principles of macro course that government determines the money supply, he laid the foundation for the policy debate about how a government operating in that fictitious world ought to go about determining the money supply.

Keynes did not arrive at a satisfactory theory of inflation, but remaining within the GT framework, the Post Keynesians did. However when Samuelsonian Macro came unstuck under the pressure of the stagflation crises of the 70's, they were swimming against a tide of received wisdom that government sets the money supply, while the Friedmaniacs were swimming with that tide.

However, one cannot lay the blame at the feet of one man for the actions of thousands of professional economists, and as Barkley Rosser, writing at EconoSpeak says:

I have been critical of Samuelson myself, and gave him quite a hard time the first time I met him in person 38 years ago. However, starting with that encounter, I must confess that while he may be the ultimate origin of much that is misguided in deeply entrenched conventional economic thought, he himself was generally personally aware of the flaws and limits of many of his own ideas, even as near the end he insisted on reasserting some of them more strongly than ever. However, I would say that the greater sins have been by simplifying followers of his, the "sons of Samuelson," who have been more responsible for codifying and spreading and enforcing the more simplistic versions of what he posited. The man himself was more complicated and self-aware than many gave him credit for , and he is indeed the last of the giant economic thinkers whose professionally active roots go back clearly into the Great Depression.

I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.