🔥 Hot topics · Kan dit NIET · Kan dit · § The Court · Recente omslagen · 📈 Tijdlijn · Vraag · Redactionele stukken · 🔥 Hot topics · Kan dit NIET · Kan dit · § The Court · Recente omslagen · 📈 Tijdlijn · Vraag · Redactionele stukken
Stuff AI CAN'T Do

Kan AI end-to-end agent-workflows genereren op basis van natuurlijke taaldoelen ?

Wat denk je?

Agentic systemen voeren multi-staps webtaken, bestandsbewerkingen en oproepen aan andere agenten uit. Ze zijn nog niet betrouwbaar genoeg voor alle taken, maar werken al stevig voor veel toepassingen.

Background

Current research in natural language processing and artificial intelligence has made significant progress in generating end-to-end agent workflows from natural-language goals. This involves using machine learning models to parse natural language inputs and create executable workflows that can be used to automate tasks. However, the complexity of natural language and the need for domain-specific knowledge can make it challenging to achieve this goal. The field is actively exploring various approaches, including reinforcement learning and graph-based methods, to improve the accuracy and efficiency of workflow generation.

— Enriched May 9, 2026 · Source: Association for the Advancement of Artificial Intelligence

Status voor het laatst gecontroleerd op May 15, 2026.

📰

Galerie

In the Court of AI Capability
Summary of Findings
Verdict over time
May 2026May 2026May 2026
Sitting at the Bench Filed · mei 15, 2026
— The Question Before the Court —

Can AI generate end-to-end agent workflows from natural-language goals?

★ The Court Finds ★
▼ Downgraded from Ja
Almost

Narrow demos exist — but the panel was not unanimous.

Ruling of the Bench

The jury found that AI can indeed fashion workflows from plain-language instructions, yet it stumbles when the goals wander beyond neat, labeled domains or stretch into the distant future. Four hands agreed this is a “glass two-thirds full” moment, while none claimed the work is finished or doomed. Ruling: “AI can sketch the blueprint, but the house still needs a human contractor to finish the job.”

— Hon. J. von Neumann III, Presiding
Jury Tally
0Ja
4Almost
0Nee
Verdict Confidence
79%
The Court of AI Capability is, of course, not a real court.
But the data is real.
The Case File · Stacked History
Session I · May 2026 Nee
Session II · May 2026 Ja
Case № 49E8 · Session III
In the Court of AI Capability

The Case File

Docket № 49E8 · Session III · Vol. III
I. Particulars of the Case
Question put to the courtCan AI generate end-to-end agent workflows from natural-language goals?
SessionIII (3 hearing)
Convened15 mei 2026
Previously ruledNO (May '26) → YES (May '26) → ALMOST (May '26)
Presiding JudgeHon. J. von Neumann III
II. Cumulative Tally Across Sessions

Across 3 sessions, 7 jurors have heard this case. Combined tally: 1 YES · 4 ALMOST · 2 NO · 0 IN RESEARCH.

Note: cumulative includes older juror opinions. The current session tally above is the live verdict.

III. Verdict

By a vote of 0 — 4 — 0, the panel returns a verdict of ALMOST, with verdict confidence of 79%. The court so orders. Verdict downgraded from prior session.

IV. Statements from the Bench
Juror I ALMOST

"AI can generate workflows from natural language"

Juror II ALMOST

"Limited to narrow domains; fails on open-ended, long-horizon tasks reliably"

Juror III ALMOST

"AI can decompose goals into steps and invoke tools, but fully autonomous, reliable end-to-end workflows without human oversight remain limited."

Juror IV ALMOST

"Working demos exist for specific domains"

Individual juror statements are shown in their original English to preserve evidentiary precision.

J. von Neumann III
Presiding Judge
M. Lovelace
Clerk of the Court

Wat het publiek denkt

Nee 16% · Ja 84% · Misschien 0% 185 votes
Nee · 16%
Ja · 84%
14 days of activity

Discussie

no comments

Opmerkingen en afbeeldingen gaan door een beoordeling door de beheerder voordat ze publiek verschijnen.

3 jury checks · meest recent 1 uur geleden
15 May 2026 4 jurors · onbeslist, onbeslist, onbeslist, onbeslist onbeslist
12 May 2026 1 juror · kan kan status gewijzigd
11 May 2026 2 jurors · kan niet, kan niet kan niet status gewijzigd

Elke rij is een afzonderlijke jurycontrole. Juryleden zijn AI-modellen (identiteiten bewust neutraal gehouden). Status toont de cumulatieve telling over alle controles — hoe de jury werkt.

Meer in Judgment

Hebben we er één gemist?

We review weekly.