ObsurfableShort answer: When the same developer-media buyer prompts are answered by ChatGPT and Perplexity, the engines disagree on the #1 outlet on 67% of questions — and they pull from different publication pools. On a 153-prompt matched panel, Perplexity mentions In Plain English on 48.4% of answers; ChatGPT mentions it on 4.6%. Perplexity puts In Plain English first on 13.7% of matched prompts; ChatGPT does so on 0.7%. On 12 of those 153 prompts, ChatGPT’s #1 is Dev.to while Perplexity’s #1 is In Plain English — same question, different “best developer blog” winner.
That gap is not a ranking tweak inside one canonical list. It is a house-favorite split between retrieval-heavy engines and ChatGPT’s training-weighted shortlist.
What we analysed
We used Obsurfable Explorer’s research corpus: buyer-style prompts tagged Technology · Developer Media — questions about where to publish technical content, which developer blogs to follow, newsletter-style outlets, and community platforms for engineers.
For each prompt we keep the latest observation per platform family (ChatGPT, Gemini, Claude, Perplexity, Grok), preferring consumer chat captures over API batch runs when both exist. We then built matched panels: prompts observed on two engines within the category window.
| Scope | Value |
|---|---|
| Corpus window | 11 March – 4 November 2026 |
| Developer Media prompts with observations | 332 |
| UI-family observations (category) | 1,143 |
| ChatGPT + Perplexity matched panel | 153 |
| ChatGPT + Claude matched panel | 148 |
Primary metrics: #1 pick (first brand extracted in the answer) and mention rate (brand appears anywhere in the ordered list). We focus on publication and community brands that dominate this category: Dev.to, Hashnode, Medium, In Plain English, freeCodeCamp, Hacker News, and Stack Overflow.
Browse live aggregates in Developer Media on Explorer, ChatGPT’s category view, and Perplexity’s category view.
What we found
1. Matched prompts rarely share the same #1 outlet
On 153 prompts observed on both ChatGPT and Perplexity:
| Metric | Value |
|---|---|
| Different #1 brand | 102 / 153 (66.7%) |
| Same #1 brand | 51 / 153 (33.3%) |
| ChatGPT #1 = Dev.to, Perplexity #1 = In Plain English | 12 / 153 (7.8%) |
| ChatGPT #1 = Dev.to, Perplexity #1 = other (not Dev.to or IPE) | 9 / 153 (5.9%) |
| ChatGPT #1 = Medium, Perplexity #1 = In Plain English | 5 / 153 (3.3%) |
The ChatGPT–Claude panel shows the same pattern at similar scale: 98 / 148 (66.2%) different #1 picks.
Dev.to is the most common #1 on both ChatGPT and Perplexity in aggregate — but second-place #1 brands diverge. Perplexity and Claude elevate In Plain English; ChatGPT elevates Medium and Hacker News more often in the #1 slot.
2. In Plain English: a search-native favorite ChatGPT under-surfaces
Mention rates on the ChatGPT + Perplexity matched panel (n=153)
| Brand | ChatGPT | Perplexity | Gap |
|---|---|---|---|
| In Plain English | 4.6% (7/153) | 48.4% (74/153) | +43.8 pp |
| Hashnode | 52.3% (80/153) | 71.9% (110/153) | +19.6 pp |
| Dev.to | 63.4% (97/153) | 75.2% (115/153) | +11.8 pp |
| freeCodeCamp | 26.1% (40/153) | 40.5% (62/153) | +14.4 pp |
| Hacker News | 29.4% (45/153) | 19.0% (29/153) | −10.5 pp |
In Plain English is the standout: it appears in roughly one of twenty-two ChatGPT answers on matched prompts, and nearly half of Perplexity answers on the same questions.
#1 pick rates on the same panel
| #1 brand | ChatGPT | Perplexity |
|---|---|---|
| Dev.to | 24.8% (38/153) | 19.6% (30/153) |
| In Plain English | 0.7% (1/153) | 13.7% (21/153) |
| Medium | 8.5% (13/153) | 1.3% (2/153) |
| Hashnode | 4.6% (7/153) | 5.9% (9/153) |
| freeCodeCamp | 5.2% (8/153) | 9.2% (14/153) |
ChatGPT + Claude matched panel (n=148) — Claude behaves like Perplexity on In Plain English:
| Brand | ChatGPT mention | Claude mention | Gap |
|---|---|---|---|
| In Plain English | 4.7% (7/148) | 43.9% (65/148) | +39.2 pp |
| Medium | 45.3% (67/148) | 33.1% (49/148) | −12.2 pp |
| #1 brand | ChatGPT | Claude |
|---|---|---|
| In Plain English | 0.7% (1/148) | 12.2% (18/148) |
| Dev.to | 23.0% (34/148) | 16.2% (24/148) |
ChatGPT still mentions Medium on 45% of matched Claude prompts — while Claude, like Perplexity, heavily favors In Plain English over Medium in both mentions and #1 slots.
3. Category-level context: Dev.to leads everywhere; the tail differs by engine
Across all 331 ChatGPT observations and 153 Perplexity observations in this category (latest-per-prompt, not only matched):
| Engine | Prompts | Top #1 outlet (share of answers) | In Plain English mention rate |
|---|---|---|---|
| ChatGPT | 331 | Dev.to — 16.6% (55/331) | 2.4% (8/331) |
| Perplexity | 153 | Dev.to — 19.6% (30/153) | 75.2% (115/153) |
| Claude | 149 | Dev.to — 16.8% (25/149) | 63.8% (95/149) |
| Gemini | 146 | Dev.to — 29.5% (43/146) | 27.4% (40/146) |
Dev.to and Hashnode rank high on every engine. The proprietary signal is the In Plain English wedge: concentrated on Perplexity and Claude, nearly absent from ChatGPT’s extracted lists on matched prompts.
4. Retrieval intensity tracks the In Plain English gap
On the 153-prompt ChatGPT–Perplexity panel:
| Engine | Web search active (share of answers) | Avg sources per answer |
|---|---|---|
| ChatGPT | 9.2% (14/153) | 2.26 |
| Perplexity | 59.5% (91/153) | 6.00 |
Perplexity answers in this panel attach sources far more often and retrieve live pages — the same conditions under which In Plain English, Hashnode, and freeCodeCamp gain mention share relative to ChatGPT. ChatGPT’s lower retrieval rate aligns with Medium and Hacker News staying in its #1 column while In Plain English stays out.
Why this is surprising
Developer-media prompts look interchangeable: “best developer blogs,” “where to publish technical tutorials,” “top engineering newsletters.” Dev.to and Hashnode appear on most engine lists — so a visibility program could assume one canonical ranking.
The matched-panel data rejects that. Two-thirds of ChatGPT–Perplexity pairs disagree on the #1 outlet. The largest mention swing is not Dev.to vs Hashnode — it is In Plain English at +44 percentage points on identical prompts. A publisher visible only on ChatGPT would miss the publication Perplexity names first on one in eight matched answers.
Limitations
- Coverage is uneven. Grok has only 10 developer-media observations in this window — too thin for matched-panel claims. Gemini and Claude have stronger samples but fewer prompts than ChatGPT.
- Brand extraction is ordered-list based. “#1” is the first extracted brand name, not a subjective “primary recommendation” in prose.
- Prompt mix spans publishing, discovery, and community questions; we report category-level panels, not a single intent slice.
- Publication brands with ambiguous names (e.g. generic “Medium”) rely on corpus normalization; In Plain English is a distinct extracted label in this dataset.
Methodology
Prompts: Active research prompts with no customer website scope, subvertical Developer Media, observed 11 March – 4 November 2026.
Platforms: ChatGPT (chat UI and API family), Gemini, Claude, Perplexity, Grok — latest observation per family per prompt; chat UI preferred over API when both exist.
Matched panels: Prompts with observations on both engines listed in each table; statistics computed on paired latest observations only.
Brands: Parsed from structured mention lists in each observation; rates are % of prompts in the panel where the brand appears (mention) or is listed first (#1).
Sources / web search: Share of answers with web search flagged active; source counts from attached citation URLs per observation.
Conclusion
In Developer Media, Dev.to remains the most common #1 across engines — but platform choice changes which publications enter the shortlist. On 153 matched ChatGPT–Perplexity prompts, In Plain English appears on 48% of Perplexity answers and 5% of ChatGPT answers; Perplexity leads with In Plain English on 14% of #1 slots versus 1% for ChatGPT. Treating ChatGPT as the whole market under-counts a publication cluster retrieval-native engines recommend on the same buyer questions — and over-weights Medium and Hacker News in the #1 column relative to Perplexity and Claude.