The Vedanta Sutras of Badarayana, Commentary by Sankara (SBE38), tr. by George Thibaut  at sacred-texts.com
18. For (the text) 'Whatever he does with knowledge' (intimates that).
In the preceding adhikarana the following conclusion has been established:--Works of permanent obligation such as the Agnihotra, if performed by a person desirous of release with a view to release, lead to the extinction of evil deeds committed, thus become a means of the purification of the mind, and thereby cause the attainment of Brahman, which leads to final release; they therefore operate towards the same effect as the knowledge of Brahman. Now the Agnihotra and similar works are either connected with a special knowledge based on the constituent members of the sacrificial work, or absolute (non-connected with such knowledge). This appears from scriptural texts such as 'He who knowing this sacrifices; he who knowing this makes an offering; he who knowing this recites; he who knowing this sings; therefore let a man make him who knows this his Brahman-priest
[paragraph continues] (Kh. Up. IV, 17, i); therefore both perform the work, he who knows this and he who does not know it' (Kh. Up. I, 1, 10).--We have now to consider the question whether only such Agnihotras and so on as are connected with knowledge cause knowledge on the part of him who desires release and thus operate towards the same effect as knowledge; or whether both kinds of works--those connected with knowledge and those not so connected--equally act in that way. The doubt concerning this point arises on the one hand from scriptural passages such as 'That Self they seek to know by sacrifice' (Bri. Up. IV, 4, 22), which represent sacrifices and the like, without difference, as auxiliary to the knowledge of the Self; and on the other hand from our observing that a superiority is conceded to Agnihotras, &c., if connected with knowledge.
Here the pûrvapakshin maintains that only such sacrificial works as are connected with knowledge are helpful towards the cognition of the Self, since we understand from various scriptural and Smriti passages that works connected with knowledge are superior to those destitute of knowledge; cp. e.g. 'On the very day on which he sacrifices on that day he overcomes death again, he who knows this' (Bri. Up. I, 5, 2); and 'Possesser of this knowledge thou wilt cast off the bonds of action;' 'Action is far inferior to concentration of mind' (Bha. Gîtâ II, 39; 49).
To this the Sûtrakâra replies, 'For what with knowledge only.' It is true that works such as the Agnihotra if joined with knowledge are superior to works destitute of knowledge, in the same way as a Brâhmana possessed of knowledge is superior to one devoid of knowledge. Nevertheless works such as the Agnihotra even if not connected with knowledge are not altogether ineffective; for certain scriptural texts declare that such works are, all of them without any difference, causes of knowledge; so e.g. the passage, 'That Self they seek to know through sacrifices.'--But, as we understand from scripture that works connected with knowledge are superior to those destitute of knowledge, we must suppose that the Agnihotra and the like if unaccompanied by knowledge are inoperative towards the
cognition of the Self!--By no means, we reply. The proper assumption is that the Agnihotra and so on, if accompanied by knowledge, possess a greater capability of originating knowledge and therefore are of superior causal efficiency with regard to the cognition of the Self; while the same works if devoid of knowledge possess no such superiority. We cannot, however, admit that the Agnihotra and similar works which scripture, without making any distinction, declares to subserve knowledge (cp. 'they seek to know through sacrifices') should not subserve it. With this our conclusion agrees the scriptural text, 'Whatever he performs with knowledge, faith, and the Upanishad that is more powerful' (Kh. Up. I, 1, 10); for this text--in speaking of the greater power of work joined with knowledge and thus proclaiming the superiority of such work with regard to its effect--intimates thereby that work destitute of knowledge possesses some power towards the same effect. By the 'power' of work we understand its capacity of effecting its purpose. We therefore accept as settled the following conclusion: All works of permanent obligation, such as the Agnihotra--whether joined with or devoid of knowledge--which have been performed before the rise of true knowledge, either in the present state of existence or a former one, by a person desirous of release with a view to release; all such works act, according to their several capacities, as means of the extinction of evil desert which obstructs the attainment of Brahman, and thus become causes of such attainment, subserving the more immediate causes such as the hearing of and reflecting on the sacred texts, faith, meditation, devotion, &c. They therefore operate towards the same effect as the knowledge of Brahman.