Here is my reconstruction of the PIE pronominal bases:

1st *me(i)
2nd *te(i)
3rd *sa

PIE used an ancient postposition *wa, originally
meaning "concerning," to mark the nominative (< ergative) forms of
these pronouns:

*me(i) wa > meu > mu
*te(i) wa > teu > tu
*sa wa > sau > so

This *wa formant was also the origin of the Proto-Semitic nominative
(< ergative).

The oblique stem of the 3rd person pronoun was suppleted by *to-
(from *ta wa).

Here is my reconstruction of the PU pronominal bases:

1st *mi
2nd *ti
3rd *sä

PU did not use *wa as an ergative/nominative formant, but it does
appear in the oblique stems of the 1sg and 2sg pronouns. This is in
accordance with the PU shift in meaning of *wa from "concerning"
to "self" (cf. Uralic reflexive > medio-passive verb formant -u).
The PU pronouns were augmented by emphatic singulative and plural
suffixes:

1sg *mi nA > minä
2sg *ti nA > tinä
3sg *sä nA > sänä

1pl *mi jA > mijä
2pl *ti jA > tijä
3pl *sä jA > säjä

Also possible is that both languages had a 1pl stem in n- (cf. Arabic
naHnu, Latin nos); whether it was inclusive or exclusive, I have no
idea. While certainly attested in the IE languages, it does not
appear to be attested in the Uralic languages, unless the Mordvin 1pl
possessive suffix -na comes from that stem.

It is my opinion that in both PIE and PU (or their parent language),
personal verb inflections came about before formalized singular and
plural pronouns. I base this on the similarity/sameness between the
plural inflections of the PIE stative and active paradigms. When it
was deemed necessary to distinguish grammatical number in pronouns,
typically a plural formant was added to the preexisting personal
suffixes (this was often the same as the plural formant for the
plural pronouns).

I believe that the Uralic personal verb inflections and possessive
suffixes have the same origin, and that they were originally simply
the early pronouns following a giving noun or verb. Tiit-Rein Viitso
writes:

"[I]n a Pre-Uralic clause, the subject usually followed the verbal
predicate. As the personal pronouns in the role of the subject were
monosyllabic and unstressed, they degenerated into suffixes, viz.
personal endings. After the rise of personal endings and personal
verb forms even predicate verbs without a personal ending are treated
as personal forms."
(http://www.eki.ee/keeljakirjandus/frameviitso.html)

I see no reason as to why that could not have been the actual state
of affairs in Pre-/Proto-Uralic. Furthermore, the common origin of
personal verb inflections and possessive suffixes suggests that Pre-
/Proto-Uralic was a mostly isolating language with little formal
distinction between nouns and verbs.

I reconstruct the Pre-/Proto-Uralic personal endings as:

1st (> 1sg) *-mI
2nd (> 2sg) *-tI
3rd (> 3sg) *-sA

1pl *-mIj(A)
2pl *-tIj(A)
3pl *-sAj(A)

- Rob