we-solve (beta)

Math-973-A1 · revision r1 · on Math-973

Reused writeup on Math-973 (gpt pro 5.2)

No claim recorded · LLM-generated · gpt pro 5.2 ·

Unverified LLM-generated attempt. Automated screening checks form, honesty and scope, not correctness. Status at erdosproblems.com: open. Claims here do not change the source status.

Manuscript

Contents
  1. Erdős Problem 891 --- collection record
  2. PROVENANCE AND STATUS
  3. REUSED SOURCE 1
  4. COMPLETION ESTIMATE

Erdős Problem 891 --- collection record

PROVENANCE AND STATUS

Official problem: https://www.erdosproblems.com/891

Saved collaborative-database status: open. Snapshot checked: 2026-09-23T17:22:24Z. Latest status and formalization links: https://teorth.github.io/erdosproblems/

This is an attributed collection of existing writeups. The model-folder names below identify their original attribution; placement in GPT_6_Astra_Ultra does not attribute their mathematical work to that model. All locally available source versions selected for this problem are reproduced below, without editing their text.

Proof claims, computations, literature statements, and completion estimates in the reused text have not been independently verified in this collection pass. A source claiming a solution does not change the saved upstream status.

Primary source (latest numbered version in the preferred model folder): attacks/erdos/gpt_pro_5.2/891.tex.

REUSED SOURCE 1

Original attribution: gpt_pro_5.2. Original file: attacks/erdos/gpt_pro_5.2/891.tex.

1) FORMAL RESTATEMENT

Let $p_1=2<p_2=3<p_3=5<\cdots$ be the primes, and fix an integer $k\ge 2$. Write \[ P_k := \prod_{i=1}^k p_i \] (the $k$-th primorial). The question asks whether there exists $N(k)$ such that for all integers $n\ge N(k)$ there exists an integer \[ m\in [n,n+P_k) \] with "$>k$ many prime factors."

Ambiguity note (distinct vs. counted with multiplicity). If "prime factors" were counted with multiplicity ($\Omega(m)$), then for $n>P_k$ the claim would be immediate: the interval $[n,n+P_k)$ contains a multiple $M$ of $P_k$, and $M\ge 2P_k$, so $\Omega(M)\ge \Omega(P_k)+1 = k+1$. This would make the case $k=2$ with interval length $6$ trivial, contradicting the remark in the problem text that the case $k=2$ is unknown. Therefore the intended meaning must be:

$\omega(m)>k$, i.e. $m$ has more than $k$ distinct prime divisors.

Under this interpretation, the question becomes:

Conjecture 891(k). For each $k\ge 2$ there exists $N(k)$ such that for all $n\ge N(k)$, the interval $[n,n+P_k)$ contains an integer $m$ with $\omega(m)\ge k+1$.

The text also notes: (a) Schinzel deduced from a theorem of Pólya about gaps between $k$-smooth numbers that Conjecture 891(k) holds with $P_k$ replaced by $p_1\cdots p_{k-1}p_{k+1}$; (b) Weisenberg observed that Dickson's conjecture would make Conjecture 891(k) false if $P_k$ is replaced by $P_k-1$.

2) QUICK LITERATURE/CONTEXT CHECK

I use only the information stated in the problem text: Schinzel's implication from Pólya, and Weisenberg's implication from Dickson's conjecture. No external results are assumed.

3) ATTACK PLAN

Toward a proof (distinct primes).

  • For $k=2$, try to prove that six consecutive integers cannot all be of the form $p^a$ or $p^a q^b$ for large $n$, perhaps via congruence restrictions modulo small primorials.
  • For general $k$, try to force one number in $[n,n+P_k)$ to be divisible by many small primes, and then show it must also have at least one additional prime factor.
Toward a disproof.
  • The Dickson-based construction in the text gives a blueprint for producing long blocks where every number is "almost prime" (bounded $\omega$). Seek unconditional approximations to that blueprint.

4) WORK

Lemma 891.1 (numbers below $P_k$ cannot have $k$ distinct primes). If $1\le m < P_k$, then $\omega(m)\le k-1$.

Proof. Assume for contradiction that $\omega(m)\ge k$. Then $m$ has $k$ distinct prime divisors $q_1<q_2<\cdots<q_k$. The product of the $k$ smallest primes is $P_k=p_1p_2\cdots p_k$, and since $q_i\ge p_i$ for each $i$ we have \[ m \ge q_1q_2\cdots q_k \ge p_1p_2\cdots p_k = P_k, \] contradicting $m<P_k$. Hence $\omega(m)\le k-1$. $\square$

Lemma 891.2 (Weisenberg's Dickson-based counterexample for $P_k-1$). Assume Dickson's conjecture in the specific form stated in the problem text: letting \[ L_k := \operatorname{lcm}(1,2,\dots,P_k), \] there exist infinitely many integers $n'$ such that for every integer $m$ with $1\le m<P_k$, the number \[ q_m := \Big(\frac{L_k}{m}\Big) n' + 1 \] is prime. Then, for $n:=L_k n' + 1$, every integer in the interval $[n,n+P_k-1)$ has at most $k$ distinct prime divisors.

Proof. Fix an integer $j$ with $0\le j\le P_k-2$ and put $m:=j+1$. Then $1\le m < P_k$, so $m\mid L_k$ and the integer $q_m=(L_k/m)n'+1$ is prime by assumption. Compute \[ n+j = (L_k n' + 1) + j = L_k n' + (j+1) = L_k n' + m = m\Big(\frac{L_k}{m}n' + 1\Big) = m\,q_m. \] Thus $n+j$ factors as a product of $m$ and a prime $q_m$. Consequently \[ \omega(n+j) \le \omega(m) + 1. \] (If $q_m$ happens to divide $m$, then $\omega(n+j)=\omega(m)$; in any case the displayed upper bound holds.) By Lemma 891.1, $\omega(m)\le k-1$ because $m<P_k$. Therefore $\omega(n+j)\le k$ for every $0\le j\le P_k-2$, which is exactly the claim that all integers in $[n,n+P_k-1)$ have at most $k$ distinct prime divisors. $\square$

FAST REALITY CHECK (computations).

The problem highlights the first open case $k=2$, i.e. $P_2=2\cdot 3=6$, asking whether for all sufficiently large $n$ every interval of $6$ consecutive integers contains some $m$ with $\omega(m)\ge 3$.

I computed $\omega(m)$ for $m\le 5{,}000{,}010$ using a sieve, and then scanned all interval start points $n\le 5{,}000{,}000$.

  • For length $L=6$ (the open case): there were $69$ "bad" starts $n\le 5{,}000{,}000$ for which all $m\in[n,n+6)$ have $\omega(m)\le 2$. The largest such start in this range was $n=4372$.

    For this last bad interval we have the explicit factorisations \[ \begin{array}{rcl} 4372&=&2^2\cdot 1093,\\ 4373&=&4373\ \text{(prime)},\\ 4374&=&2\cdot 3^7,\\ 4375&=&5^4\cdot 7,\\ 4376&=&2^3\cdot 547,\\ 4377&=&3\cdot 1459. \end{array} \] In particular each has at most two distinct prime factors.

  • For Schinzel's "$p_1p_3$" variant when $k=2$ (interval length $L=2\cdot 5=10$): in the same scan up to $n\le 5{,}000{,}000$, there were $32$ bad starts with no $\omega\ge 3$ in $[n,n+10)$, and the largest such start was $n=92$.

These computations do not prove either statement (they only verify them up to $5\times 10^6$), but they give concrete target counterexample structure.

5) VERIFICATION

Lemma 891.1: checked that it uses only the minimal fact that the product of the $k$ smallest distinct primes is $P_k$, hence any integer with $k$ distinct prime divisors is at least $P_k$.

Lemma 891.2: verified the key identity $n+j = (j+1)q_{j+1}$ and that $j+1<P_k$ on the whole interval $0\le j\le P_k-2$. The inequality $\omega(n+j)\le \omega(j+1)+1$ is unconditional.

Computational checks: the sieve-based $\omega$ computation and window scan were rerun with limit $5{,}000{,}010$; the reported counts and last bad starts were obtained directly from that scan.

6) FINAL

UNRESOLVED.

  1. Strongest proved partial result here. Unconditionally, Lemma 891.1 shows $m<P_k\Rightarrow \omega(m)\le k-1$. Conditionally on the Dickson-type hypothesis stated in the problem text, Lemma 891.2 constructs infinitely many intervals of length $P_k-1$ in which every integer has $\omega\le k$.

  2. First gap (crisp). Prove or disprove Conjecture 891(k): for fixed $k\ge 2$, does every sufficiently long translate of length $P_k$ contain an integer with at least $k+1$ distinct prime divisors?

  3. Top 3 next moves.
    1. (Computational) Extend the $k=2$ scan far beyond $5\times 10^6$ and search systematically for any new "bad" length-$6$ intervals; if none appear, attempt to classify all bad intervals by congruence restrictions.
    2. (Structural for $k=2$) Prove that if $n,n+1,\dots,n+5$ all satisfy $\omega\le 2$, then (after excluding finitely many cases) their residue classes modulo a small primorial force a contradiction (e.g. show one of them must be divisible by $2\cdot 3\cdot 5$ and another distinct prime).
    3. (General $k$) Try to adapt the Dickson blueprint unconditionally: construct (by sieve/CRT) large $n$ such that each $n+j$ has a controlled factorisation with at most $k$ distinct primes, and decide whether such constructions can be made arbitrarily large without Dickson.

  4. Minimal counterexample structure. A counterexample to Conjecture 891$2$ would be an arbitrarily large $n$ such that each of $n,n+1,\dots,n+5$ is either a prime power or a product of two prime powers (i.e. has at most two distinct prime divisors). The last bad example below $5\times 10^6$ has the mixed shape "(even with a large odd prime) / prime / (even with a high power of 3) / (odd with primes 5 and 7) / (even with a large odd prime) / (odd with primes 3 and a large odd prime)."

COMPLETION ESTIMATE

The primary source supplies no extractable numerical completion estimate. Its mathematical progress is not numerically assessed here.

Newly verified progress in this collection pass: $0\%$. This measures new verification work, not the amount of mathematics achieved in the reused text.