Only the best are using them...
Credibility score: 48/100 — Mixed Credibility. Several questionable claims detected. Watch with healthy skepticism.
BSmeter analyzed "Only the best are using them..." and rated it 48/100 for credibility (a BS score of 52/100 — mixed credibility), on 2026-06-09. Its weakest claim — "Only top 0.01% of engineers use loop engineering due to massive token budgets" — scored 35/100 and was flagged as dubious. 14 claims were checked against the video transcript. Scores are produced by BSmeter's AI analysis of the transcript, not independent human verification.
Of 14 claims analyzed: 2 scored under 40, 12 between 40 and 69, and 0 at 70 or above.
Claims analyzed
Video opens with a highlight reel preview — Just Vibes (50/100)
At 0:00
Teaser promises a brand-new 'loops' meta that everyone suddenly discovered this weekend
Why this score: Just Vibes segment. No specific claim to verify yet — this is pure hype setup for the Boris clip and Peter's viral tweet that follows.
Original quote: “A new coding meta just dropped. Over this past weekend, everybody started talking about loops. The two main characters in the world of AI coding talked about it at the same time. Peter Steinberger and Boris Churnney. Here is an interview from Boris that went absolutely viral. >> I don't prompt…”
Monthly reminder not to prompt agents directly anymore — Opinion (50/100)
At 0:30
Monthly reminder framing a workflow shift as established fact.
Why this score: Opinion on best practice.
- Positions "loops" as the new standard without evidence of widespread adoption
- "Monthly reminder" implies consensus that doesn't exist yet
- Community comments call this rebranded old techniques
Original quote: “Here's your monthly reminder that you shouldn't be prompting coding agents anymore. You should be designing loops that prompt your agents. ... this video is sponsored by here.”
Only three types of AI loop triggers exist — Dubious (45/100)
At 4:30
Three categories feels tidy but the 'human kicks it off' bucket swallows half of real usage.
Why this score: Oversimplification.
- Real systems mix webhook events, cron, manual runs, and agent-initiated loops
- The speaker's own later example (agent keeps looping until goal) doesn't fit neatly into any single bucket
- Community comments already flag this as rebranding old automation patterns
Original quote: “All of these other ones are based on certain actions that happen. So there really are only three total types of triggers. One, some kind of action happens like a PR opens. Two, it's a schedule that happens, so every 30 minutes. And three, a human kicks it off.”
Here.now launched private storage recently — Unverifiable (50/100)
At 6:30
No date, no announcement — just 'recently' with zero receipts.
Why this score: Unverifiable because no source or date is provided.
- 'Recently' is too vague to check against any timeline.
- Web search shows no matching product announcement for Here.now private storage.
Original quote: “they also recently launched private storage. So you don't need to always just publish everything publicly, you can have your agent store pretty much anything on here.now.”
Here.now added custom domain support even more recently — Unverifiable (50/100)
At 6:42
Same problem — 'even more recently' and no link to verify.
Why this score: No evidence in current search results for Here.now custom URLs.
- Claim is too recent and unsourced to confirm or deny.
- Absence of proof ≠ proof of absence, so score stays neutral.
Original quote: “And then even more recently, they launched custom URLs. So rather than only having a here URL, you can use your custom domain with here. Now publish directly to it.”
Here.now is completely free right now — Unverifiable (50/100)
At 6:52
Free today? Cool. No pricing page or terms cited though.
Why this score: Pricing claim left hanging without source.
- Could be a limited promo or genuinely free tier.
- No web results confirm current pricing status.
Original quote: “And the best part, it's completely free right now.”
Cloud Code has built-in /loop command for recurring prompts — Unverifiable (50/100)
At 7:08
Sounds handy, but no official docs or screenshots shown.
Why this score: Feature existence cannot be confirmed from transcript alone.
- No external link or changelog reference provided.
- Commenters note similar looping already exists in other tools.
Original quote: “And they literally have a feature called /loop. So, you just start typing /loop and it says, "Run a prompt or slash command on a recurring interview." So, loop 5 minutes and then whatever you want.”
Looping every 5 minutes will rack up a trillion-dollar token bill — Opinion (50/100)
At 7:32
Hyperbolic joke, not a real cost projection.
Why this score: Clear exaggeration for comedic effect.
- No actual token pricing or usage math is shown.
- Functions as rhetorical warning rather than literal forecast.
Original quote: “Obviously, that's ridiculous and I'm going to have a trillion dollar token bill at the end of the month, but that's how you do it.”
Only top 0.01% of engineers use loop engineering due to massive token budgets — Dubious (35/100)
At 8:30
0.01% is a hell of a flex — no receipts, just vibe-based hierarchy of who can afford the bill.
Why this score: No public data backs the exact percentile.
- Speaker names Peter Steinberger and Boris from OpenAI/Anthropic as examples
- High token budgets exist at frontier labs, but "1% of 1%" is pulled from thin air
- Community comments already call this rebranded old techniques
Original quote: “only the top 1% of 1% are using these techniques like loop engineering because not only are they enabled to try these new techniques but also they're given infinite or very high token budgets which only a few companies in the entire world can really afford”
OpenAI and Anthropic give staff unlimited tokens — Unverifiable (50/100)
At 10:30
No public info on either company's internal token policy — just taking his word.
Why this score: Unverifiable internal policy claim.
- No company announcements or leaks confirm unlimited tokens
- Plausible given their compute access, but zero external evidence
- Speaker presents it as established fact without sourcing
Original quote: “Both companies give their employees infinite tokens. That's why they're able to experiment.”
Peter Steinberger had $1.3M monthly token bill — Unverifiable (50/100)
At 10:37
Sounds dramatic, but no receipts or link provided — just dropped the number.
Why this score: No supporting link or date given.
- Speaker says "a few weeks ago" with zero citation
- $1.3M is an enormous figure; would expect public discussion if real
- Could be real private usage, but currently uncheckable
Original quote: “Peter Steinberger showed about a few weeks ago he had $1.3 million in monthly token usage.”
AI agent loops are the inevitable future of engineering — Opinion (50/100)
At 10:49
Bold prediction dressed as certainty — future is rarely that clean.
Why this score: Opinion presented as inevitability.
- No evidence this specific model will dominate
- Assumes linear progress with no plateaus or pushback
- Many past "this is the future" tech claims aged poorly
Original quote: “this is absolutely the future of engineering. We will continue to build the software factory that builds the software.”
Questions whether humans stay in AI loops forever — Opinion (50/100)
At 12:03
Philosophical musing, not a testable claim.
Why this score: Speculative framing.
- Poses a question rather than asserting a fact
- Future human role in AI systems is genuinely open
- No data exists to settle this yet
Original quote: “Will the human be required in the loop forever or not?”
AI will soon design its own factories and companies via recursive self-improvement — Dubious (35/100)
At 12:30
Calls this the 'recursive self-improvement' threshold — Anthropic's essay exists but no factory-designing AI is running today.
Why this score: No current system is autonomously designing factories or companies.
- Recursive self-improvement remains a theoretical milestone, not an achieved capability.
- Anthropic published discussion pieces, but no deployed AI is closing the loop on physical infrastructure.
- The jump from code generation to industrial self-replication is still science fiction.
Original quote: “imagine in which I am no longer setting the direction. I am no longer setting the goal. AI has taste and it's able to decide what features to build, what products to build, what companies to build. And really what we're describing is AI is able to design its own factory. That is the point at which…”
See the full analysis with timestamps →