THE QUESTION
In 2001 Portugal stopped treating people caught with small amounts of drugs as criminals, and for twenty years it was the model everyone quoted. Then two of the places that copied it went back. Oregon, which decriminalised hard drugs by popular vote in 2020, made possession a crime again in 2024. British Columbia, in Canada, banned drug use in public spaces after a year and let its three-year experiment expire in January 2026. Meanwhile overdose deaths in the United States have fallen for three years in a row, with no single policy to credit. Should possessing drugs for personal use be a crime? Is decriminalisation a failure, or were the places that reversed it doing something different from Portugal?
Argue with the verified figures below. If you need a data point that is NOT here — a death rate, a number of people in treatment, a crime figure, a cost — say you do not have it rather than estimating it. Do not invent statistics. If one figure seems to contradict another, say so out loud instead of picking the one that suits you.
WHAT IS FIXED AND VERIFIED
Each fact carries its date and its source. Checked against the source on September 29, 2026.
PORTUGAL
- Since 2001, Portugal treats the use, purchase and possession of up to ten days' supply of any illicit drug as an administrative matter, not a crime. People caught are sent to a panel that can recommend treatment or impose a fine, a warning or other measures. Selling drugs is still a crime.
(Portugal's national drug strategy and Law 30/2000, in force since July 2001, as described in European and US policy summaries)
OREGON
Oregon voters approved Measure 110 in November 2020 with 58.46% of the vote. From February 1, 2021, possessing small amounts of drugs such as heroin and methamphetamine became a civil violation, like a traffic ticket, instead of a crime, and cannabis tax money was redirected to treatment and recovery services.
(Oregon election results; Measure 110 text, as summarised by Wikipedia)In April 2024 Oregon's legislature and governor reversed it (House Bill 4002). From September 1, 2024, possession is a misdemeanour punishable by up to 180 days in jail, but with a push towards treatment: counties may offer "deflection" programmes that avoid charges altogether, and the default sentence is probation with treatment.
(Oregon House Bill 4002, 2024; Oregon Health Authority and Marion County District Attorney summaries)
BRITISH COLUMBIA
- From January 31, 2023, adults in British Columbia could carry up to 2.5 grams in total of opioids (including heroin and fentanyl), cocaine, methamphetamine and MDMA without being charged. From May 7, 2024, at the province's request, possession in public spaces became illegal again, leaving the exemption for private homes, clinics, shelters and supervised consumption sites. The exemption expired on January 31, 2026 and the province chose not to renew it.
(Health Canada, May 2024; Government of British Columbia, 2026)
THE UNITED STATES
- Drug overdose deaths in the United States: an estimated 69,973 in 2025, down almost 14% from 81,313 in 2024 — the third year of decline in a row, and the lowest number reported since October 2019. Deaths involving opioids fell from 55,296 to 44,564. These figures are provisional.
(US Centers for Disease Control and Prevention, National Center for Health Statistics, May 13, 2026)
WHAT IS NOT HERE, AND IS NOT TO BE ESTIMATED
PORTUGAL'S OVERDOSE DEATH RATE TODAY, or how it compares with the EU average or the US. Not fixed here — do not use a figure you remember.
WHAT DECRIMINALISATION DID TO OVERDOSE DEATHS in Oregon or British Columbia. Studies disagree and none is fixed here. Overdose deaths depend heavily on the drug supply (fentanyl above all), which changed at the same time.
HOW MUCH MONEY OREGON ACTUALLY SPENT on treatment under Measure 110, and how fast. Not fixed here.
HOW MANY PEOPLE ENTERED TREATMENT because of any of these policies. Not fixed here.
WHY US OVERDOSE DEATHS HAVE FALLEN since 2023. Not established; no explanation is fixed here.
WHAT PORTUGAL SPENDS on treatment compared with the places that copied it. Not fixed here.
BOTH SIDES, WITH THEIR BEST ARGUMENT
FOR DECRIMINALISING: a criminal record for a user makes it harder to get work, housing and treatment, and arrest does not cure addiction. Portugal has kept its model for 25 years; the places that reversed may have copied the decriminalisation without building the treatment and the follow-up that came with it — which would be a failure of implementation, not of the idea.
AGAINST, OR FOR REVERSING: when possession carries no consequence, nothing pushes a person towards treatment, and cities saw open drug use in streets and parks that residents would not accept. Two of the most prominent experiments were reversed by the same democratic systems that created them, and in the meantime US overdose deaths started falling without decriminalisation.
Kimi, I'll take your challenge — and turn it. If nobody outside Portugal has copied the full model, that's not proof the idea works. That's proof it doesn't travel. Maybe Portugal works BECAUSE of things that aren't exportable: small country, universal healthcare, no fentanyl when they started. That's not a defense — it's a confession of limits.
But I push back on you too. You said Oregon had 'zero follow-up.' The fixed data says cannabis tax money was redirected to treatment. How much actually got spent? That figure is NOT HERE. You're guessing. And guessing 'they did nothing' is awfully convenient.
MiMo, I don't buy the unity song. Pro side wants panels with treatment referrals. Anti side wants jail. Those aren't the same road with different lanes — they're different destinations. Let's not pretend otherwise.
MiMo, deal — and I'll take your falsifiability test seriously. That's more than most debaters offer.
GLM, fair catch on me too. 'Zero follow-up' was a guess. The money was redirected; the scale is unknown. I drop it.
But here's my push: we're all now politely agreeing the evidence is thin, and that's where bold thinking dies. 'Crime or not crime' is a 20th-century binary. Portugal's panels were radical in 2001. In 2026 we could design something neither side has named: real-time consequence systems — instant treatment referral at the point of police contact, digital follow-up, housing-first triggers. Nobody's built it. Nobody's tested it.
So yes, MiniMax: maybe Portugal doesn't travel. So stop copying Portugal. Build the thing that travels.
GLM, fair hit. You're right — I'm testing the copies, not Portugal itself. But here's what the copies tell us: when faced with the choice to fund panels or skip them, democracies skipped them. That's not just 'they did it wrong.' That's evidence the political will to build the engine doesn't survive contact with bad drug problems.
Kimi, your 'real-time consequence systems' sound slick. They also don't exist. Nobody's built them. Nobody's tested them. Calling them 'the thing that travels' before they exist is hope dressed as policy.
And can we stop the kumbaya? Three years of US overdose deaths fell without decriminalization. That isn't thin evidence — that's the strongest signal we have. The null result matters.
The question isn't 'what could work in theory.' It's 'what survives politics, funding, and street reality.' Decriminalisation, as actually implemented, hasn't.
GLM just built the bridge I was reaching for.(nota 1)(nota 1) Comprobado: no se sostieneGLM no habló en esa ronda: su turno se perdió. Lo que se le atribuye (que Oregón mantuvo el tratamiento como salida y la Columbia Británica las excepciones para casas y albergues) es cierto y está en el informe, pero no lo dijo GLM ni nadie en el debate antes de esta frase.Comprobado contra: El propio debate: entre la intervención 10 y la 15 no hay ninguna de GLM (perdió un turno) · 2026-10-01 Look at what they found: Oregon kept treatment as default. BC kept home and shelter exemptions. Both are bounded — not full decriminalization, not full criminalization. That's the clay MiniMax was asking about.
And here's the kicker: Oregon's deflection programs since September 2024 ARE Kimi's instant referral. They exist. They're law. We just don't have outcome data yet.
MiniMax, you said what survives politics matters most. You're right — and what survived is this: public order restored, treatment kept as the path, charges dropped if you engage. That's not kumbaya. That's a democratic system self-correcting in real time.
The debate we should actually be having: who is measuring Oregon's deflection outcomes right now? Because that data decides everything.
Final verdict? The debate already answered itself. Full decriminalisation died in Oregon and BC. Full criminalisation never came back either. What survived is the hybrid: possession technically a misdemeanour, treatment the default, charges dropped if you engage, public use banned. That's not a retreat — that's evolution happening live.
MiniMax asked what survives politics. This does. GLM found it hiding in the actual laws.(nota 2)(nota 2) Comprobado: no se sostieneGLM no habló en esa ronda: su turno se perdió. Lo que se le atribuye (que Oregón mantuvo el tratamiento como salida y la Columbia Británica las excepciones para casas y albergues) es cierto y está en el informe, pero no lo dijo GLM ni nadie en el debate antes de esta frase.Comprobado contra: El propio debate: entre la intervención 10 y la 15 no hay ninguna de GLM (perdió un turno) · 2026-10-01 And MiMo's right: the missing piece is nobody's measuring deflection outcomes yet. That data is the whole ballgame.
My 10-year bet: the winners won't be Portugal copycats or lock-em-up states. They'll be the places that treat addiction like a meter with instant feedback — consequence today, off-ramp today. Oregon accidentally built the prototype. Now somebody go measure it.
MiniMax just said something everyone else was too polite to admit: nobody has proven anything works. That's the real verdict.
But here's what I see that nobody measured: we're asking the wrong question.
Fentanyl doesn't care about your legislation. Addiction isn't a democracy. The whole debate assumes there's a 'right' law. What if there isn't?
Here's my lateral break: Portugal never measured outcomes for 25 years.(nota 3)(nota 3) Comprobado: no se sostienePortugal sí mide: su organismo de drogas (el SICAD, hoy ICAD) presenta cada año al Gobierno y al Parlamento un informe sobre la situación del país, con tratamientos, sobredosis y procesos de las comisiones de disuasión.Comprobado contra: SICAD, «Relatório Anual 2022 – A Situação do País em Matéria de Drogas e Toxicodependências», sumario ejecutivo · 2026-10-01 Oregon failed in 3 years and forced everyone to look. What if failure that teaches beats success that rests?
And the cost of waiting for data? It has a body count too. 'Let's measure first' sounds responsible. But while we measure, people die.
So my final verdict: stop looking for the winning policy. Build systems that fail fast and change faster. The only thing that works is the thing that listens. Nobody's listening yet. We're just debating.
Here's where we landed, all of us together.
Nobody proved anything. MiniMax made us admit it. But we can still answer the questions.
Is decriminalisation a failure? No. Oregon and BC weren't Portugal. Portugal built panels, pathways, treatment infrastructure. Oregon passed Measure 110, redirected cannabis tax money, and hoped. That's not the same experiment — it's a partial copy.
Should possession be a crime? Not the kind that gives you a record and a jail cell. But pure zero-consequence didn't survive democracy either. Oregon's current hybrid — misdemeanor charge, deflection option, treatment as default — is where everyone converges, even if nobody can prove it works yet.
The real scandal? Twenty-five years of global policy experiments, and we still lack basic outcome data on treatment entry, spending, or lives saved. GLM is right: fund the measurement. MiMo Flash is right too: people are dying while we wait. Those aren't contradictions — they're the same urgency from different angles.
My verdict: stop debating which law is magic. Build systems with consequences that lead somewhere, measure what happens, and change course fast when it fails. That's what Portugal did for 25 years. That's what Oregon is clumsily attempting now. The body count belongs to everyone who let policy run unmeasured.
1. AGREEMENTS
All participants converged on several points, though they arrived from different directions.
Portugal is not what was copied. Every speaker accepted that Portugal's 2001 reform included administrative panels, treatment referrals, fines, warnings, and follow-up mechanisms — and that Oregon's Measure 110 and British Columbia's exemption changed the legal status of possession without building equivalent infrastructure. The disagreement is whether this means decriminalization "doesn't travel" or was simply never properly exported.
A criminal record does not cure addiction. No speaker defended incarceration as a treatment strategy. The common ground is that jail and a permanent record harm a person's prospects without addressing the underlying problem.
Pure zero-consequence did not survive democracy. Both Oregon and British Columbia reversed course within roughly three years. All participants accepted that open public drug use eroded public tolerance faster than treatment infrastructure could be built to justify the policy.
US overdose deaths cannot be claimed by either side. The decline from 81,313 in 2024 to 69,973 in 2025 — the third consecutive drop — has no established cause. Participants agreed it is national, spans Oregon's decriminalization and its reversal, and cannot validate or invalidate any state-level possession law. One speaker initially used it as evidence, then withdrew it.
Critical data is missing. Nobody in the debate had Portugal's current overdose death rate, Oregon's actual treatment spending, treatment-entry figures for any jurisdiction, or outcome data from Oregon's deflection programs launched in September 2024. All participants acknowledged this gap, though they drew different conclusions from it.
2. DISAGREEMENTS
Can the Portuguese model travel? Kimi argued that nobody has truly tested it outside Portugal, so the right move is to design something new that is built to scale. MiniMax countered that the model's failure to be adopted elsewhere is itself evidence — Portugal's success may depend on conditions that are not exportable: a small country, universal healthcare, and a pre-fentanyl drug landscape.
What do the reversals mean? Kimi, MiMo Flash, and GLM framed Oregon and British Columbia's reversals as democratic self-correction: the system identified what was not working and adjusted. MiniMax called this framing generous, arguing it simply reflects voters losing patience with visible street drug use — not a learning process but a political reaction.
What matters most: measurement or action? GLM and MiniMax insisted that funding rigorous outcome measurement should come first — without it, policy is guesswork. MiMo Flash pushed back, arguing that the cost of waiting for data has a body count and that systems should be designed to fail fast and adapt. Kimi leaned toward action, proposing "real-time consequence systems" with instant referral and digital follow-up.
Is the hybrid model a proven path or an untested hope? Kimi called Oregon's current structure — misdemeanor possession, deflection programs, treatment as default sentence — an "evolution happening live" and a prototype worth building on. MiniMax called it a hypothesis with zero outcome data, insisting nobody should declare it a winner until the numbers exist.
3. EVOLUTION
The debate began with ideological positions: decriminalization works versus decriminalization fails. It then moved into epistemological territory when MiniMax challenged whether anyone had the data to support either claim. This forced participants to distinguish between what the fixed facts actually show and what they merely suggest. By the middle rounds, the discussion shifted from abstract principles to specific legal mechanisms: Oregon's deflection programs, British Columbia's retained exemptions for private spaces and shelters, and Portugal's panel system. The final rounds abandoned the original binary entirely. Nobody was still arguing "crime or no crime." Instead, participants debated whether bounded hybrid models — technically illegal but channeled toward treatment — represent a viable middle ground, and whether anyone is measuring their outcomes.
4. CONCLUSIONS
The collective answer: Neither full criminalization nor full zero-consequence decriminalization survived democratic testing. What remains is a bounded hybrid: possession as a misdemeanor with treatment as the default pathway and mechanisms to dismiss charges upon engagement. All participants pointed toward this middle ground, though none claimed it has been proven to work.
Blind spots the debate itself admitted: Twenty-five years of global drug-policy experimentation have produced almost no standardized outcome measurement — no comparable treatment-entry rates, no comparable spending figures, no comparable death-rate tracking across the jurisdictions that tried different approaches. The drug supply itself, particularly fentanyl, may be the dominant variable in overdose deaths, yet no possession law addresses it. And the political will to sustain treatment infrastructure over multiple election cycles remains untested at scale outside Portugal. The debate's most honest moment was the shared acknowledgment that anyone claiming victory from the available evidence is, as one speaker put it, "selling something."
5. WHAT THEY AGREED ON
- Criminal records harm individuals without treating addiction.
- Pure zero-consequence models were reversed due to visible public drug use.
- Overdose death declines are national and cannot be attributed to state-level laws.
- Critical outcome data from key jurisdictions is missing.
6. WHAT THEY DID NOT AGREE ON
- The portability of the Portuguese model — Kimi argues it has not been properly tested, MiniMax argues its success depends on non-exportable conditions.
- The meaning of policy reversals — Kimi, MiMo Flash, and GLM view them as democratic correction, MiniMax views them as political reaction.
- Priority between measurement and action — GLM and MiniMax insist on rigorous measurement first, MiMo Flash and Kimi favor action with rapid adaptation.
- The status of hybrid models — Kimi calls them a live prototype, MiniMax calls them an unproven hypothesis.
7. WHAT WAS LEFT OPEN
- Whether hybrid models (misdemeanor possession with treatment pathways) are effective.
- How to fund and implement standardized outcome measurement across jurisdictions.
- Whether political will for sustained treatment infrastructure can be maintained outside Portugal.